Synopsis: Ict: Data: Data:

The available data indicates that while the FDI sector is of vital importance to the growth of exports

Subdivision of data into North & South Tipperary areas not available 28 Case study: Benefi ts of participation in EU Projects

-refining & Bioenergy, Data Analytics and Manufacturing Research. Two Research Centres are hosted by the University of Limerick:

c) Data pertaining to employment, turnover and exports in the Region from the Central Statistics

Baseline Data: The initial sections of this Strategy form a baseline assessment of the current research

•Priority Area B-Data Analytics, Management, Security & Privacy •Priority Area C-Digital Platforms, Content & Applications

the lack of up-to-date data, in particular in developing countries. ITU is joining the international statistical community in looking into ways of using new and emerging data sources †such as those

associated with big data †to better provide timely and relevant evidence for policy-making. Calls for

huge amounts of data, as well as being a major source of big data in their own right. Big data from

I trust that the data and analysis contained in this report will be of great value to the ITU membership

The 2014 edition of the Measuring the Information Society Report was prepared by the ICT Data and

and Michael Minges to the compilation of data on international bandwidth, revenue and investment. Helpful inputs and suggestions were received

The report includes data from Eurostat, OECD, IMF, Informa, the UNESCO Institute for Statistics, the

ITU also appreciates the cooperation of countries that have provided data included in this report The report was edited by Anthony Pitt and Bruce Granger, ITU English Translation Section.

5. 3 Telecommunication data and their potential for big data analytics...181 5. 4 Big data from mobile telecommunications for development and for better monitoring...

Annex 2. ICT price data methodology...231 Annex 3. Statistical tables of indicators used to compute the IDI...

5. 2 An overview of telecom network data...182 5. 3 Customer profiling using telecom big data...

1. 4 What is a data revolution?..30 2. 1 ITU discussion forums on ICT statistics...

5. 3 How mobile operators currently use data to track service uptake, business performance and

5. 4 Using mobile data for development...187 5. 5 How mobile network data can track population displacements †an example from the

2010 Haiti earthquake...188 5. 6 Leveraging mobile network data for transportation and urban planning in Sri lanka...

189 5. 7 Poverty mapping in CÃ'te d†Ivoire using mobile network data...190 5. 8 Using mobile-phone data to track the creditworthiness of the unbanked...

191 5. 9 Using mobile big data and mobile networks for implementing surveys...193 List of tables

1. 1 Rural population covered by a mobile-cellular signal, 2012.4 1. 2 Total Internet domain registrations by world region, 2003,2008 and 2013.18

The data also show a continuous increase in Internet usage, with growth in the number of

and data) market segments, and considering both subscriptions and household access data. This will be followed by a presentation of the latest

trends in terms of investment and revenue in the telecom sector. Then, a number of key indicators

WSIS+10 review, the demand for a data revolution and the role of big data for ICT monitoring

data reveal, however, that digital divides still exist and that some people are excluded still from access to communication networks

For countries where data are available, the number of mobile subscriptions far exceeds the number of mobile phone users (Partnership, 2014

many developing countries for which data are available, but is closing with the availability of

Further research and data would be necessary to determine people†s access to, and use of, voice

Partnership (2014) based on ITU data Overall mobile-cellular population coverage %Rural population covered %Rural population

The data on fixed-and mobile-broadband uptake confirm what has been observed on the ground. In developed countries, fixed

data based on ITU and Telecom Advisory Services calculations more and more countries upgrade their mobile networks. As mentioned earlier, 2g population

Data on 3g population coverage are less available. According to ITU estimates, global 3g population coverage stood at around 50 per cent

New data collected by ITU on the deployment of fibre transmission capacity in countries shows

at the data also reveals major disparities across regions: Asia and the Pacific (in particular China

Partnership (2014) based on ITU data Percentage of rural population covered by at least a 3g mobile

delivering data-intensive applications and services through high-speed networks. While fibre transmission networks constitute an

The latest ITU data show that by end 2014, almost 44 per cent of the world†s households will have

data are available, rural household access falls far below urban household access, with differences ranging from 4 per cent (meaning

2014). 8 Available data also show that Internet access in rural households is growing slowly

but data are not readily available for those countries. As has been illustrated earlier, network deployment is still

Data on public access is collected by ITU through its household questionnaire, but only few countries

report data on this indicator, let alone broken down by urban and rural populations. Public access can be provided by commercial facilities

The data include both global top-level domain (gtld) and country code top-level domain (cctld) registrations

Data sourced from Facebook 0 500 1'000 1'500 2013 2012 2011 2010 2009

International data on ICT access and use by enterprises are collected annually by the United Nations Conference on Trade and Development

In the developing world, data on ICTS in enterprises are scarce and only collected by few

Data supplied by Zooknic, compiled from cctld and other sources. Figures exclude fifteen cctlds which act as virtual

There is little data on the use of ICTS by government organizations, and those countries that do have data are usually the more advanced

ones with high levels of connectivity in general More information is available about government services provided online, tracked by the United

The latest data show that, today, governments of all countries have established central websites and that more

Data from United nations E-government Survey (2014 Chart 1. 21: E-government services provided by countries (transactional services, left, and e-participation

Data from United nations E-government Survey (2014 101 73 60 46 44 42 41 40

exchange of forms, data and other information Internet access in schools also achieves cost efficiencies by automating manual tasks and

The latest available data from the UNESCO Institute for Statistics (UIS) 22 show that, in

developed countries for which data are available In developing countries, school access to Internet is lower on average, although much

Data on broadband in secondary schools in Bangladesh are not available. Data for Nicaragua, Philippines and Indonesia do not include upper

secondary. Data for European countries and Costa rica refer to lower secondary. Data for Guyana, Nicaragua and Indonesia refer to primary and lower

secondary. Data for Cambodia include pre-primary schools. Data for Morocco, Tunisia, Guyana, Montserrat, Dominican republic, Nicaragua, Colombia

Trinidad and tobago, Bangladesh, Philippines, Sri lanka, Azerbaijan, Bhutan, Cambodia, Kazakhstan, Malaysia, Maldives, Singapore, Belarus and the Russian Federation refer to public schools.

In Suriname, there are no private schools in upper secondary. Data for Palestine refer to West bank schools

only Source: UIS database, Partnership on Measuring ICT for Development WSIS Targets Questionnaire, 2013 P

e rc e n ta g e o f s c h o o ls

Data for Philippines refer to primary and lower secondary. Data for Venezuela refer to primary only.

Data for Montserrat and Saint lucia refer to secondary only. Data for Palestine refer to West bank schools only.

Data for Bahrain, Belarus, Morocco and Tunisia refer to 2008. For Morocco, ICT -qualified teachers figures refer to 2008.

Data for Azerbaijan, Barbados, Jordan, Saint lucia, Singapore, Trinidad and tobago, Uruguay, Philippines and Sri Lanka refer to public schools only

Source: UIS database, Partnership Questionnaire on WSIS Indicators, 2014 Proportion of teachers trained to teach basic computer skills (or computing

%P ro p o rt io n o f te a c h e rs

Available data collected by UIS at the international level shows that education systems in countries seem to put more emphasis on

document also includes two targets on data and monitoring and stresses the â€oeneed to take urgent

availability of disaggregated data to ensure that no one is left behindâ€. 26 The role of ICTS as a key development enabler

collect ICT data at the national level Monitoring progress towards achievement of the WSIS outcomes has been an integral component

indicators and collecting data and statistics on ICT. Since its creation in 2004, the Partnership has

national-level data on the indicators identified to measure the WSIS targets. The results of the survey are featured in the report

along with other data sources The Final review of the WSIS targets: Achievements, challenges and the way forward report is available at:

Using different data sources and big data analytics An important element in the discussion related

to the post-2015 development agenda and the setting of measurable goals and targets has

data and statistics are collected, analysed and disseminated, in view of the large data gaps prevailing in many developing countries in basic

statistics in the areas of the economy, health education, labour, etc. all of which are crucial to

see Box 1. 4 on the data revolution. Since then it has received considerable attention within the

development data According to the discussions, a data revolution considers new data sources, in addition to existing official sources:

besides governments other stakeholders such as the private sector, civil 27 Measuring the Information Society Report 2014

Box 1. 2: A decade of successful international cooperation on ICT measurement In 2004, ICT measurement was still in its infancy, with little

calls surfaced for reliable and comparable data in order to take stock of the emerging information society, identify digital

of their national ICT data collection. The methodological work developed by the Partnership has contributed significantly to

Data availability has increased also significantly over the past decade. In particular, there are more comparable data on ICT

infrastructure, household access and Internet users. For example at the beginning of the century, only around a dozen developing countries collected data on Internet users, while today there are

almost 50 developing countries collecting this indicator through official surveys (Chart Box 1. 2). Data on household access to

the Internet or a computer are now being collected by more than 100 economies worldwide, and data on Internet use in

businesses by almost 70 countries, although not on a regular basis (Partnership UNSC 2011. Similarly, whereas no data were

available on ICT access and use in schools, they have started now to be compiled in many developing countries (see section 1. 5

At the same time, major data gaps remain, in particular in developing countries and LDCS. This concerns, notably, statistics

public-sector organizations, ICT-related employment data, as well as data related to online security and cybercrime, gender

and youth, and cultural and environmental aspects. The growing information society will increasingly require more and better

Internet user data through official surveys, by level of development Note: Chart shows countries that have collected data on the number of

Internet users through official national surveys. Data are presented in three-year intervals and include countries that have collected

data for at least one year within those intervals Source: ITU 0 2 11 24 36

49 1 3 27 37 41 40 1 5 38 61 77 89 0 20

New data sources could include big data mostly provided by private-sector companies which could help â€oeimprove the timeliness and

completeness of data, without compromising the relevance, impartiality and methodological soundness of the statistics†(UNSC, 2014

recognized that â€oebig data constitute a source of information that cannot be ignored by official statisticians†and that â€oeofficial statisticians must

contributions from Member States. c Due to data limitations, currently mobile-broadband signal coverage is considering in determining this target

d Data being compiled by the Global Cybersecurity Index (GCI Source: ITU 29 Measuring the Information Society Report 2014

data and legislation related to big data •To address the issue of obtaining â€oeaccess at no cost†to big data from the private sector

issue of access to transborder data or access to data on transboundary phenomena •To develop guidelines to classify the

various types of big data sources and approaches •To develop methodological guidelines related to big data, including guidelines

strategy for data providers and users on the issue of use of big data for official statistics

in IT issues or in the use of open data platforms The UN Global Working group on Big data for

data for policy applications; and build public trust in the use of private-sector big data for official

ICTS are part of the debate on the data revolution big data and, more broadly, emerging data issues

in the post-2015 development debate. First the ICT sector in itself represents a new source

of data, provided by, for example, Internet and telecommunication companies. Second, the spread and use of ICTS allow public and private

store and analyse huge amounts of data. At the same time, however, monitoring access to and

can be used as an alternative data source. Without ICTS, no ICT-driven data revolution will take place In view of the link between big data and ICTS, work

is under way in ITU with a view to contributing to the debate and identifying new ways and means

of big data, in particular data coming from ICT companies; that regulatory authorities should explore the development of guidelines on how

•Which techniques are needed for data anonymization for aggregated datasets such as mobile-phone records •How is exploited big data in different

What is a data revolution The report of the High-level Panel of Eminent Persons on the

has called for, inter alia, a data revolution taking advantage of new technology and improved connectivity: â€oewe also call for a data revolution for sustainable

development, with a new international initiative to improve the quality of statistics and information available to people and

levels on what such a data revolution could entail and how it could be implemented. While no internationally agreed concept

part of a data revolution: 32 •In view of the ubiquitous availability of communication networks, the use of new information technologies (e g

the collection and dissemination of data •Data should be further disaggregated (by gender income, age, geography, etc.

to ensure that no one is left out; in this regard, traditional statistical processes should be made more efficient

•The focus should go beyond data dissemination and also include investment in the development of concepts

•New, nontraditional data sources should be explored and leveraged to complement existing ones and satisfy the demand for data needs in new areas, such

as big data, geospatial information and geographical information systems •Open data policies should be envisaged to ensure

accountability and promote transparency •The role of data, statistics and monitoring for policy -making and decision-making should be increased

31 Measuring the Information Society Report 2014 •What is the link between big data and open data (crowdsourcing, cloud

computing, etc •is there need a to regulate data management and service providers •How can market dominance in the area

of big data be prevented and the rights of the data owners protected ICT data collection and analysis

•How can big data complement existing ICT statistics to better monitor information-society developments •Which type of data from ICT companies

are most useful and for which purposes •Which new ICT indicators could be produced from big data sources

•What are the key issues that need to be addressed, and by whom, in terms of

complement official ICT data •How can big data from telecommunications inform not only ICT but broader development policy in

5 Data collection on Europe and North america will follow in 2014 6 For a list of IXPS, see for instance http://www. datacentermap. com/ixps. html

19 Data refer mostly to the year 2011 20 The UN E-government Development Index is a composite benchmarking indicator based on a direct assessment of the state

elements that should be part of a data revolution 33 See final report of WTIS-13, available at:

better data become available. For example what was considered basic infrastructure in the past †such as fixed-telephone lines †is fast

•Data availability and quality. Data are required for a large number of countries as the IDI is a global index.

There is relative paucity of ICT-related data, especially on ICT usage, in the majority of developing

countries. In particular, the three indicators included in the skills sub-index should be considered as proxies until

data directly relating to ICT skills become available for more countries •The results of various statistical analyses.

data and to explore whether the different dimensions are statistically well-balanced While the basic methodology has remained the

related data availability, the indicators included in the IDI and its sub-indices are under regular

field of ICT statistics and data collection, work through online discussion forums and annual face-to-face meetings.

data with data based on national household surveys (demand-side indicators. An indicator such as the percentage of individuals using a

reported these data to ITU for at least one year between 2011 and 2013. It is therefore too

subscription data in the IDI with mobile-phone user data. In view of the methodological difficulties in collecting harmonized data on

international Internet bandwidth, a review of the definition of the indicator is currently under discussion in EGTI

The IDI was computed using the same methodology as in the past, applying the following steps (Figure 2. 2 and Annex 1

•Preparation of the complete data set. This step includes filling in missing values using various statistical techniques

•Normalization of data. This is necessary in order to transform the values of the IDI indicators into the same unit of

•Rescaling of data. The data were rescaled on a scale from 0 to 10 in order to

compare the values of the indicators and the sub-indices •Weighting of indicators and sub-indices.

This chapter presents the IDI based on data from 2013 in comparison with 2012. It should be

•Country data revisions. As more accurate data become available, countries provide ITU with revised statistics for

previous years, which have been taken into consideration. This also allows ITU to identify inconsistencies and revise

added based on data availability. Overall this version of the IDI includes 166 countries/economies as compared with

According to data from the European union EU), 85 per cent of Danes have some level of

Data also show that the Republic of korea achieves the highest advertised fixed-broadband speeds, with all

domestic demand for data driven by the high volume of local content, and domestic Internet

Data from the EU confirm that household access is highly correlated with regular use of the Internet, which

Data from the European Commission†s Digital Agenda underline the competitiveness of the European fixed

mobile data plans. 35 During 2013, operators further extended their wireless infrastructure and services throughout Thailand,

Based on the 2013 and 2012 data presented in this chapter, the current (2013) global divide is

Data change very little over time and advances in skills do not show immediate effects

Data not available 59 Measuring the Information Society Report 2014 the profitability of various kinds of economic

selected for their high data availability for a large number of countries Past editions of this report (see MIS 2013) have

Data on urban population, population density, land area and GNI p. c. are sourced from the World bank.

Data on population size are sourced from UNPD.**Correlation is significant at 0. 01 level

pronounced when it comes to data on Internet access and use. Access to the Internet (be it

data are available for both sets of indicators. The following steps were performed •First, developed countries were excluded

in respect of which data pertaining to both the MDG indicators and the IDI are available for at least 16 countries. 41 The MDG

147 economies for which data were available for 2002 and 2013 shows that the global IDI value has doubled almost from 2. 52

would require different sets of data (including micro data) collected from official surveys Therefore, the analysis should be considered as a

first step in an attempt to quantify the relationship between ICT performance and MDG progress. The

data are available •Almost all MDG indicators that are included under MDG 1, MDG 4, MDG

children, through the analysis of health data collected through public health applications and by serving as a platform for exchange and

of smartphones to capture essential data on the patients and monitor their treatment has accelerated progress.

of MDG indicators, where 2002 and 2011 data are available for both sets of indicators.

data are available for 2002 and 2011 for both the MDG indicators and the IDI

earlier section where 2011 data were analysed improvements in the level of ICT access and use between the ten-year periods have shown

analysis will require different data sets, including micro data on ICT usage collected from official

surveys. Micro data offer analysts and researchers ample information and considerable flexibility to apply quantitative models that identify

relationships and interactions between indicators and topics covered in a survey, thereby fostering the diversity and quality of research and analysis

2 Data on the indicators included in the skills sub-index are sourced from the UNESCO Institute for Statistics (UIS.

http://data. worldbank. org/about/country-classifications/country-and-lending-groups#High income 7 Based on 2011 data:

https://ec. europa. eu/digital-agenda/sites/digital-agenda/files/DAE%20scoreboard%202013%20-%20

19 In these countries, the in-scope population for data on Internet users is aged individuals 16-74

as well as the household access data, excludes transient labourers, which account for a significant proportion of residents in Qatar.

According to data from ICTQATAR, â€oetransient labourers†make up 27 per cent of the overall population 33 http://qnbn. qa/qatar-vision-2030

35 http://www. telecompaper. com/news/thai-operators-reduce-prices-of-smartphone-data-plans--900198

World bank, see http://data. worldbank. org/indicator/NY. GNP. PCAP. CD 39 See: Czernich, N.,Falck, O.,Kretschmer, T. and Woessmann, L. 2009), Broadband Infrastructure and Economic growth

if there is available data for at least one year between 2010 and 2012. The number of countries included in the analysis varies from 16 to 101 developing countries.

since it is the closest year to available MDG data 42 Broadband Commission. Transformative Solutions for 2015 and Beyond a Report of the Broadband Commission Task force

Data from South African operators show that not only is wireless -broadband penetration reaching higher levels

consuming more data, indicating an increase in the intensity of usage. MTN reported a growth

of 63 per cent in data volumes in the first half of 2013 and Vodacom reported that on average

users were generating 75 per cent more data traffic per device than a year ago. 5 Wireless

Data from household surveys show that the actual number of people using a mobile-cellular phone is

Data from the annual ICT household survey show that, since 2008, computers have replaced telephones as

Data from household surveys collected in a number of CIS countries underline that mobile -cellular penetration, measured as the number

Data from the Eurobarometer underlines this finding on average, 92 per cent of European union citizens (the majority of countries in the region

Data on Individuals using the Internet for Eurostat members are sourced from Eurostat. Eurostat collects data for Internet users aged 16-74 years old

Source: ITU World Telecommunication/ICT Indicators database World Developed 0 10 20 30 40 50

coverage, and the majority of all data traffic is carried by the LTE network. 27 Very high increases

11 Internet user data from Gulf countries are not comparable, as they refer to different populations.

Data from Bahrain and Qatar refer to the overall population, i e. including expatriate/transient workers.

Data from United arab emirates are estimated by ITU based on base data excluding the transient worker population 12 Reported in activated external capacity

13 http://www. ofca. gov. hk/en/industry focus/telecommunications/facility based/infrastructures/submarine cables/index html 14 http://submarinenetworks. com/systems/intra-asia/sjc/sjc-cable-system

/26 Data reported by the country refer to 2012 27 http://www. verizonwireless. com/wcms/consumer/4g lte. html and http://www. telecompaper. com/news/verizon-wireless

data following a harmonized methodology since 2008. Initially, prices were collected for fixed -telephone, mobile-cellular (voice and SMS) and

Since 2012, the data collection has been extended to include mobile -broadband prices. These data have proved to be

useful for the international comparison of ICT prices across more than 160 countries, and for identifying those cases where prices constitute a

They include end-2013 data for each of the three price sets contained in the IPB (fixed-telephone

Based on 140 economies for which 2008-2013 data on fixed-telephone and mobile-cellular prices were available

-broadband service with a 500 MB monthly data allowance. 7 Despite the limitations of comparing

included in the ITU price data collection exercise Growth rates have been stagnating since 2008 and there have been few structural changes in

price data had already fully or partially liberalized their fixed-telephone market in 2008, compared

Based on 140 economies for which 2008-2013 data on fixed-telephone and mobile-cellular prices were available

*Data correspond to the GNI per capita (Atlas method) in 2013 or latest available year adjusted with the international inflation rates.**

ranked because data on GNI p. c. are not available for the last five years

GNI p. c. and PPP$ values are based on World bank data Rank Economy Fixed-telephone sub-basket GNI p. c

*Data correspond to the GNI per capita (Atlas method) in 2013 or latest available year adjusted with the international inflation rates.**

ranked because data on GNI p. c. are not available for the last five years

GNI p. c. and PPP$ values are based on World bank data Rank Economy Mobile-cellular sub-basket GNI p. c

total mobile data traffic was offloaded onto fixed networks in 2013 (CISCO, 2014), highlighting the role that fixed broadband plays in supporting the

which 2008-2013 data on fixed-broadband prices were available Source: ITU 0 50 100

and data on fixed (wired)- broadband subscriptions by speed, which show that a significant share

Based on 143 economies for which 2008-2013 data on fixed-broadband prices were available Source:

Based on 165 economies for which 2013 data on fixed-broadband prices were available Source: ITU

GNI p. c. values are based on World bank data 0. 0 0. 5 1. 0 1. 5

GNI p. c. values are based on World bank data 0 2 4 6 8 10 12

GNI p. c. values are based on World bank data for fixed-broadband uptake in the country.

GNI p. c. values are based on World bank data 0 5 10 15 20 A s

GNI p. c. values are based on World bank data There are 13 countries in the Asia and the Pacific

Indeed, the latest data on international connectivity show that this may remain an issue in Kiribati (45 Mbit/s), Marshall

GNI p. c. values are based on World bank data However, a comparison with other regions shows

*Data correspond to the GNI per capita (Atlas method) in 2013 or latest available year adjusted with the international inflation rates.**

ranked because data on GNI p. c. are not available for the last five years

GNI p. c. and PPP$ values are based on World bank data Rank Economy Fixed-broadband sub-basket

which ITU collects data on prices (Figure 4. 1 operators are adding plans for specific devices

allowing customers to pool the data consumed by different devices in a single subscription. 23

with neither time nor data volume constraints This is the common scheme in a vast majority

include data volume caps, e g. USD 10 for 50 MB per month. Several operators also offer

of time and data volume limitations, e g. USD 5 for one day of use with a maximum of 50 MB

priced on the basis of the data allowance (i e. the data volume in MB included in the plan) and not

the speed. Many operators do not even advertise the speed of the mobile-broadband service, but

in data allowances, bundled voice minutes and SMS, time limitations, premium speeds, etc rather than actual differences in prices for the

data allowance was PPP$ 36.6 (or USD 24.4) for prepaid plans and PPP$ 30.0 (or USD 19.2) for

-broadband service with 500 MB monthly data allowance was PPP$ 25.3 (or USD 16.9) for

the monthly data allowance was half as large Nevertheless, the reduction in price was not

proportional to the reduction in the data allowance, confirming that the price per GB is

lower for larger data allowances, the equivalent of a volume discount. Unlike in the case of

Based on 119 economies for which data on mobile -broadband prices were available for the four types of plans

data allowance of 500 MB are about eight times more affordable in developed countries than in developing countries, on average (Chart

service with 1 GB monthly data allowance corresponded to more than 20 per cent of GNI

Based on 119 economies for which data on mobile -broadband prices were available for the four types of plans

Based on 119 economies for which data on mobile -broadband prices were available for the four types of plans

much less than 500 MB of Internet data per month, supported by the fact that several African

limited data allowances, and even Internet radio would need to be limited This suggests that, if mobile broadband is to

Based on 119 economies for which data on mobile -broadband prices were available for the four types of plans

Country data for The americas reveal that there are a number of countries which have high prepaid computer

include a monthly data allowance much larger than 1 GB (Table 4. 8). This is the case of Chile

Such high monthly data allowances for prepaid mobile-broadband dongles suggest that these services target high-end customers

data allowances in The americas (Table 3. 7 suggesting that postpaid rather than prepaid is the base offer for regular computer-based

-broadband plans with a monthly data allowance of 1 GB suggest that mobile broadband could be

Percentages are calculated on the basis of the total number of countries with data available in each region:

data allowance of at least 1 GB. Although the minimum data allowance is the same, in practice

most fixed-broadband plans allow unlimited data use (Table 4. 4), whereas most computer -based mobile-broadband plans with a minimum

monthly data allowance of 1 GB really do have a cap of 1 GB (Tables 4. 7 and 4. 8

In almost half of the African countries included in the price benchmark, mobile-broadband prices were more than USD 10 cheaper per

*Monthly data allowance (MB as%of GNI p. c. USD PPP$ 1 Austria 0. 13 5. 31 4. 62 48'590 1'024

*Monthly data allowance (MB as%of GNI p. c. USD PPP$ 80 Maldives 2. 08 9. 7 12.6 5'600 700

*Data correspond to the GNI per capita (Atlas method) in 2013 or latest available year adjusted with the international inflation rates.**

ranked because data on GNI p. c. are not available for the last five years

GNI p. c. and PPP$ values are based on World bank data Chapter 4. ICT prices and the role of competition

*Monthly data allowance (MB as%of GNI p. c. USD PPP$ 1 Norway 0. 1 8. 34 5 102'610 500

*Data correspond to the GNI per capita (Atlas method) in 2013 or latest available year adjusted with the international inflation rates.**

ranked because data on GNI p. c. are not available for the last five years

GNI p. c. and PPP$ values are based on World bank data Rank Economy Mobile-broadband prepaid handset-based (500 MB

*Monthly data allowance (MB as%of GNI p. c. USD PPP$ 80 Antigua & Barbuda 2. 58 27.78 33.8 12'910 1'024

*Monthly data allowance (MB as%of GNI p. c. USD PPP$ 1 Austria 0. 13 5. 31 4. 62 48†590 1

*Data correspond to the GNI per capita (Atlas method) in 2013 or latest available year adjusted with the international inflation rates.**

ranked because data on GNI p. c. are not available for the last five years

GNI p. c. and PPP$ values are based on World bank data Rank Economy Mobile-broadband postpaid computer-based (1 GB

*Monthly data allowance (MB as%of GNI p. c. USD PPP$ 76 Montenegro 2. 89 17.49 28.93 7†260 3

*Monthly data allowance (MB as%of GNI p. c. USD PPP$ 1 Austria 0. 13 5. 31 4. 62 48†590 1

*Data correspond to the GNI per capita (Atlas method) in 2013 or latest available year adjusted with the international inflation rates.**

ranked because data on GNI p. c. are not available for the last five years

GNI p. c. and PPP$ values are based on World bank data Rank Economy Mobile-broadband prepaid computer-based (1 GB

*Monthly data allowance (MB as%of GNI p. c. USD PPP$ 76 South africa 4. 82 28.90 51.17 7†190 1

to making available more data on household and individual economic welfare, as well as its distribution. This section presents a refined

data on the distribution of household income or consumption expenditure are available The objective is to explore how factors such as

Data on household income, on the other hand measure only people†s economic welfare, and do not include the business sector.

Data are collected by national statistical offices by means of household income and expenditure surveys HIES) or household surveys including a module

classify income data by deciles, individuals are placed in ascending order according to the household income attributed to them, and then

availability of products, may be relevant. 34 Data on household consumption expenditure are also obtained from household surveys

international comparisons based on data on household economic welfare, because there are several methodological issues that limit

-level economic data provide information (not available from macroeconomic indicators) on the actual income and expenditure capacity

previously mentioned, these data can be used to obtain a finer-grain indication of the affordability

Data for the United states and Sweden are sourced from the OECD Database on Income Distribution

Data for South africa and Viet Nam are sourced from the World Bank†s Povcalnet and refer to 2008

4. 18 uses data on income inequality to reveal differences in the affordability of fixed-broadband

Household disposable income based on World Bank†s Povcalnet data adjusted with ITU estimates on average persons per household

inequality within countries, economic data at the household level also make it possible to determine more precisely the affordability

Household consumption based on World Bank†s Povcalnet data adjusted with ITU estimates on average persons per household

and expenditure data, the following conclusions can be drawn •In more than 85 per cent of countries

for which data are available, the richest 20 per cent of the population can afford

data are available, a basic fixed -broadband subscription still represents more than 5 per cent of household

countries for which data on household income or expenditure distribution are available, fixed-broadband plans remain

Data on household disposable income refer to 2011 or latest year available.**†Lowest 20%†refers to the price divided by the average income of the first

Household disposable income for OECD countries and the Russian Federation based on data from the OECD Database on Income Distribution

on World Bank†s Povcalnet data adjusted with ITU estimates on average persons per household Fixed-broadband prices as a%of household disposable income%households

Data on household consumption expenditure refer to 2011 or latest year available.**†Lowest 20%†refers to the price divided by the average expenditure

Household consumption expenditure based on World Bank†s Povcalnet data adjusted with ITU estimates on average persons per household

limited by lack of data on the distribution of household income/expenditure Available data suggest that basic fixed

-broadband plans are affordable for 90 per cent of the population in Jordan and Tunisia, whereas they are less affordable in

with a 500 MB monthly data allowance) as a percentage of disposable household income and

collects price data (Chart 4. 13), and is currently the mobile-broadband service that is available in most

which data are available. This is also the situation in Armenia, Dominican republic and Egypt In these countries, a prepaid handset-based

subscription data show that, in developed countries, handset-based mobile-broadband subscriptions are individual, rather than shared

Data on household disposable income refer to 2011 or latest year available.**†Lowest 20%†refers to the price divided by the average income of the first

Household disposable income for OECD countries and the Russian Federation based on data from the OECD Database on Income Distribution

data adjusted with ITU estimates on average persons per household Prepaid handset-based mobile-broadband prices

Data on household consumption expenditure refer to 2011 or latest year available.**†Lowest 20%†refers to the price divided by the average expenditure

Household consumption expenditure based on World Bank†s Povcalnet data adjusted with ITU estimates on average persons per household

Data show that handset-based mobile -broadband prices (with a 500 MB monthly allowance) are affordable for a majority of the

Data on equivalized household income and consumption expenditure per decile refer 2011 or latest available year

Equivalized household disposable income for OECD countries and the Russian Federation based on data from the OECD Database on Income

Equivalized household disposable income and consumption for other countries based on World Bank†s Povcalnet data adjusted with ITU

Data for the Arab States are limited to four countries. In Tunisia and Jordan, individual handset-based mobile-broadband subscriptions

On the basis of the data presented, it can be concluded that income inequality does not only

of comprehensive data series on the prices for these two services, which makes it possible to study

data for developing countries. Quantitative studies are in several cases restricted to samples of EU and OECD countries, or else

regulatory data for up to 144 countries in the period 2008-2013. Including such a large set

on telecommunication data from EU, OECD and specific countries hold true in a global context

the level of detail of the indicators, since data availability and comparability for such a large

Simple averages for 140 economies with available data on fixed-broadband prices and competition for the period 2008-2013

Herfindahl-Hirschman Index (HHI) data sourced from Informa Chart 4. 22: Competition in mobile markets and mobile

Simple averages for 140 economies with available data on mobile-cellular prices and competition for the period 2008-2013

Herfindahl-Hirschman Index (HHI) data sourced from Informa competition and the openness of the market to private and foreign investment

based on panel data regression. This enables us to go beyond descriptive statistics and draw some robust conclusions on the link between

analyses in this section, data from the Regulatory Tracker have been extracted for clusters 1, 2 and 3, and combined into a single value per country

for up to 144 countries based on data for the five-year period from 2008 to 2013

Panel data regression is a statistical technique which is used to assess how variations in a set

GB of data allowance included per month However, entry-level fixed-broadband plans in several countries offer higher

speeds and larger data allowances. In order to measure how these enhanced features affect prices, a variable on fixed

capped data allowances are included as controls in the model for fixed-broadband prices Chapter 4. ICT prices and the role of competition

Data collected by ITU, see Annex 2 for more details on the methodology for the collection of

Data collected by ITU, see Annex 2 for more details on the methodology for the

Data sourced from Informa Herfindahl-Hirschman Index for mobile cellular (voice and SMS with Li being the number of mobile-cellular subscriptions of firm i,

Data sourced from Informa 159 Measuring the Information Society Report 2014 Box 4. 2: Panel regression models for fixed-broadband and mobile-cellular prices (continued

Data collected by ITU, see www. itu. int/tracker for more information Fixed-broadband speed:

Speed of the entry-level fixed broadband plan in Mbit/s. Data collected by ITU

if the fixed-broadband plan includes a cap in the monthly data allowance, and 0 otherwise.

Data collected by ITU Descriptive statistics of the dependent variables Average Standard deviation Minimum Maximum 2008 2013 2008 2013 2008 2013 2008 2013

Descriptive statistics calculated for 124 economies that have complete data for the two models Source: ITU

Fixed-broadband plans with data caps are linked to prices 31 %lower than unlimited plans Fixed-broadband speed 0. 016

data caps by fixed-broadband service providers is correlated with cheaper entry-level fixed -broadband plans, other things being equal.

ITU data collection considers a minimum of 1 GB Chapter 4. ICT prices and the role of competition

data cap of 1 GB or more, or allows unlimited traffic. In 2013, entry-level fixed-broadband

plans included a data cap in 35 per cent of the countries included in the fixed-broadband

data caps were enforced in entry-level fixed -broadband plans. This suggests that operators can offer cheaper prices in exchange for reduced

data consumption, thus indicating that capacity in fixed-broadband networks is still an issue in several countries,

additional Internet data beyond 1 GB is still non -negligible in many countries Finally, different entry-level fixed-broadband

sector, such as operators†strategies on data caps competition in the fixed-broadband market and the ICT regulatory environment, may together be

is the existence of data caps, which is correlated with lower prices. This may indicate that fixed-broadband capacity

*Data correspond to the GNI per capita (Atlas method) in 2013 or latest available year adjusted with the international inflation rates

GNI p. c. and PPP$ values are based on World bank data Chapter 4. ICT prices and the role of competition

data caps, competition in the fixed -broadband market and the ICT regulatory environment, are together more of

information on the PPP methodology and data, see http://icp. worldbank. org 3 GNI takes into account all production in the domestic economy (i e.

http://data. worldbank. org/indicator/NY. GNP. PCAP. CD 4 Voice over internet services, such as Skype or Voipbuster, are excluded from the analysis in this section

9 Based on 2012 and 2011 ITU data for countries accounting for 97 per cent of global fixed (wired)- broadband subscriptions, it

http://data. worldbank. org/about/country-classifications/country-and-lending-groups 11 The Communications Commission of Kenya (CCK) issued mobile virtual network operator licences to three operators in April

These data suggest there is limited competition among operators 13 Advanced mobile technologies refer to standards agreed by the ITU Radiocommunication Assembly for next-generation

http://www. verizonwireless. com/wcms/consumer/shop/shop-data-plans/more-everything. html 24 The details of the different †4g†plans offered by Tigo can be found on the following websites

25 Data for mobile-broadband services have been collected since 2012 through the ITU ICT Price Basket Questionnaire,

http://data. worldbank. org/about/country-classifications/country-and-lending-groups 31 Household incomes include wages, salaries, self employment incomes, capital and property income, private transfers and

35 Differences in the equivalence scales of the source data used in this chapter are corrected roughly using ITU estimates on

World bank Povcalnet data on income distribution are published per capita, and figures from the OECD Database on Income Distribution are equivalized using the square root scale

According to 2008/9 survey data, mean household consumption in Angola was almost the same as in

38 Data on income distribution are aggregated per household and then equally attributed to each member of the household

In addition, data on income distribution are averaged per decile. If the price of a fixed-broadband plan represents less

In conclusion, the data presented do not determine whether 100 per cent of a population within a country can afford a fixed-broadband plan,

The World Bank†s Povcalnet data are adjusted using ITU estimates on the average number of inhabitants per household

market share data were available. This includes 95 economies from the developing world and 44 from the developed world

economies for which price and market share data were available for 2013. This includes 99 economies from the developing

share data were available. This includes 96 economies from the developing world and 44 from the developed world

-to date and reliable data, in particular from developing countries. The information and communication technologies (ICT) sector is

of those statistics and identify new data sources In this context, the emergence of big data holds

existing, but often limited, ICT data There is no unique definition of the relatively new phenomenon known as big data.

most basic level it is understood as being data sets whose volume, velocity or variety is very

data flows originating from, among other things commercial transactions, private and public records that companies and governments collect

data source. While existing data can provide a relatively accurate picture of the spread of

telecommunication networks and services there are significant data gaps when it comes to understanding the development of the

information society. Relatively little information for example, is available on the demand side While an increasing number of countries

currently collect data on the individual use of ICTS, many developing countries do not produce such information (collected through

Consequently, not enough data are available about the types of activity that the Internet is

public services, even fewer data are available to show developments over time and enable informed policy decisions.

In addition to the data produced and held by telecommunication operators, the broader ICT sector, which includes not just telecommunication

captures a wide array of behavioural data Together, these data sources hold great promise for ICT monitoring,

and this chapter will explore the potential of today†s hyper-connected digital world to expand on existing access and

data with global socioeconomic coverage. In particular, mobile telephone access is quasi -ubiquitous, and ITU estimates that by the end of

data for development, highlight advances, point to some best practices and identify challenges including in regard to the production and

and type of data that telecommunication companies, in particular mobile-cellular operators, produce, and how those data are

175 Measuring the Information Society Report 2014 currently being used to track ICT developments and improve their business.

data may be used to complement official ICT statistics and assist in the provision of new evidence for a host of policy domains

Administrative data •Electronic medical records •Insurance records •Tax records Commercial transactions •Bank transactions (inter-bank as well as personal

increasing amounts of data from different data sources. Indeed, one of the key trends fostering

data. One of the earliest definitions, introduced by the Gartner consultancy firm, describes big data characteristics such as velocity and

which data are generated, assessed and analysed, while the Chapter 5. The role of big data for ICT monitoring and for development

term â€oevariety†encompasses the fact that data can exist as different media (text, audio and

relating to data quality and provenance, and the uncertainty that may exist in their analysis (IBM

data (Jones, 2012)( Figure 5. 1 Included within the scope of big data is the category of transaction-generated data (TGD), 2

also sometimes described as â€oedata exhaust†or â€oetrace dataâ€. These are digital records or traces

Most of the data captured by telecommunication companies can be classified as TGD As is often the case with technological

data deluge. Encouraged by promising results but also reduced budgets, the public sector is turning towards big data to improve its service

of data generated through large-scale datafication and digitization of information Different types and forms of

data, including large amounts of unstructured data Level of quality accuracy and uncertainty of data and data

sources VALUE Potential of big data for socio -economic development VELOCITY Speed at which data are

generated and analyzed 177 Measuring the Information Society Report 2014 a specific condition that then led Walmart to

improve its production chain †in this case, by increasing the supply of Pop Tarts to areas likely

to be affected by a disaster. Walmart has also made use of predictive analytics, which uses personal information and purchasing patterns

proficient in their use of data-driven decision making have been found to have productivity levels up to 6 per cent higher than firms making

minimal to no use of data for decision-making Brynjolfsson, Hitt and Kim, 2011. Significantly industries now have the ability to conduct

unstructured data, Vestas was able to significantly improve customer turbine location models and optimize turbine performance

weather and location data in ways that were previously not possible. These new insights have led to improved decisions

These fundamental shifts in data exploitation to generate new socioeconomic value coupled with the simultaneous emergence

of new rich data sources that can potentially be linked together and analysed with ease have sparked also the interest of governments

data analytics to a range of activities, including monitoring and improvement of tax compliance and revenues, crime detection and prediction

data they collect and generate themselves Chapter 5. The role of big data for ICT monitoring and for development

leveraging data from new sources, including crowd-sourced data generated by the public In the United states, for example, Boston City

Hall released the mobile app â€oestreet Bump†which uses a phone†s accelerometer to detect

richest data sources for enabling governments and development agencies to improve service delivery are actually external.

data include those captured and/or collected by the private sector, as well as the digital breadcrumbs left behind by citizens as they go

and bigâ data analytics to improve public administration, from land management to the administration of benefits.

the data captured by the use of ICTS. This broadly includes data captured directly by telecommunication operators as well

as by Internet companies and by content providers such as Google, Facebook, Twitter etc. Big data from the ICT services industry

data, including big data from the ICT industry is subject to national regulation. In the EU, for

example, a number of directives require data producers to obtain users†consent before gathering any of their personal data. 5

data in action and of the great potential of big data for broader development and monitoring

Mayer-Schã nberger and Cukier, 2013; Mcafee and Brynjolfsson, 2012. GFT worked by monitoring health-seeking behaviour expressed

focusing on the use of search-engine data to understand dengue fever outbreaks, 6 monitor prescription drug use (Simmering, Polgreen and

data beyond the realm of user search terms Online job-posting data are being used to

supplement traditional labour statistics in the United States7 and other countries. In another effort, an academic project at MIT known as the

data from hundreds of online retailers. 8 The data are used then by researchers to understand a

action, has been mining Twitter data from Indonesia (where Twitter usage is high) 9 to understand food price crises.

analytics on the Twitter data to forecast the consumer price index several weeks in advance Byrne, 2013. As discussions on the post-2015

Pulse is also using Twitter data to understand and compare the relevance of different development topics among countries (Box 5. 2

Qos) data on broadband quality. For example the United states Federal Communications Commission (FCC) has released mobile apps that enable consumers to check their

Mobile data Despite the rapid growth in Internet access 60 per cent of the world†s population is still

Depending on the source of Internet data results may also be biased more or less. A 2013 study into the characteristics and behaviour

is very similar to real life (census) data on the average age when American people get

Mobile data are already being utilized for research and policy-making, not only in developed but also in developing economies

economic developments. 13 Data are also being used to improve responsiveness in the event of natural disasters or disease outbreaks.

data to understand the spread of malaria in Kenya (Pindolia et al. 2012; Wesolowski et al 2012a), and of cholera in Haiti after the 2010

data from the ICT sector, and especially those available to telecommunication operators have wide applicability for informing multiple

Leveraging such data to complement official statistics and facilitate broader development will enable governments as well as development agencies to better

better understand the type of data that can be made available 5. 3 Telecommunication data

and their potential for big data analytics Fixed and mobile telecommunication network operators, including Internet service providers

ISPS), are an important source of data and for the purpose of this chapter, all forms of

telecommunication big data (either volume velocity or variety) are being considered. Most telecommunication data can be considered as

TGD, 14 that is, the result of an action undertaken such as making a call, sending an SMS,

-cellular service, data from mobile operators have the greatest potential to produce representative results and reveal developmental insights

Telecommunication data The mobile telecommunication data that operators possess can be classified into different types, depending on the nature of the information

they produce. They include traffic data, service access detail records, location and movement data, device characteristics, customer details and

tariff data. For a more detailed overview of these types of data, see Chapter 5 Annex

To collect traffic data, operators use a range of metrics to understand and manage the traffic

flowing through their networks, including the measurement of Internet data volumes, call, SMS and MMS volumes, and value-added service

VAS) volumes. Internet service providers can also use deep packet inspection (DPI), 15 which is a special process for scanning data packages

transiting the network Service access detail records, including call detail records (CDRS), are collected by operators

whenever clients use a service. They are used to manage the infrastructure and for billing purposes, and include information on the

time and duration of services used and the technology used, for example, for the mobile network (2g, 3g, etc..

These data are potentially also very useful for building a rich profile of customers, as outlined in this section

positioning data, with the latter providing more detailed and precise location information Since mobile user devices used to access mobile

An overview of telecom network data Source: ITU, adapted from Naef et al. 2014 Tr affi

Data volume Call volume SMS/MMS volume Erlang DPI data Timestamp of use Contact network

Duration of use Applicable charges Handset type Technology utilized (2g, 3g, DSL/ADSL, etc Billing address

Passive positioning data e g. cell ID IMEI Active positioning data e g. cell triangulation, GPS MAC address

Customer demographics (e g. age, gender, national ID card number Billing address Payment history (postpaid Billing address

Transaction generated data Stored warehouse data 183 Measuring the Information Society Report 2014 subscriber†s purchasing power (see below for

Finally, operators maintain tariff data in the form of billing records for their current and past

How mobile operators currently use data to track service uptake, business performance and revenues Operators use their TGD to monitor the uptake and penetration

On the basis of the detailed service-usage data collected telecommunication operators can produce a range of detailed

of originating and terminating calls, SMS and MMS usage, data upload volumes, data download volumes, level of use of different

VAS, and level of use of different OTT services. These data can be reported as averages (over time or for different categories of

user), as well as at various levels of aggregation (again over time or for different categories of user.

Finally, service consumption data are used to produce revenue data and projections at various levels of disaggregation or

aggregation. For example, the average revenue per user (ARPU) is a KPI for operators, which identify their most important customers

will, for instance, often associate revenue data with resource allocation to ensure that Qos at the base stations used by their

others, use aggregated revenue data to track and benchmark countries†ICT developments, monitor the evolution of the information society and

lines, for example when selling data Customer profiling Telecom operators capture a range of behavioural data about their customers

Chapter 5. The role of big data for ICT monitoring and for development 184 Customer profiles include details about

data regulations governing the manner in which operators may keep and/or use such data.

This being the case, the extent to which behavioural profiling is used by operators may vary greatly

also seek to monetize the data they hold. The simplest way of doing this is to sell (anonymized

data to third parties. The customer insights obtained through the analysis of usage data can also help create new business lines, either

through innovation (e g. new types of VAS) or by partnering with other businesses, including credit-scoring and related financial services

Cignifi, 19 which obtains data from mobile operators and financial institutions to build credit profiles and evaluate customer creditworthiness

on existing and new sources of data for the post-2015 development agenda (United nations 2013).

a report on â€oebig data and modernization of Chapter 5. The role of big data for ICT monitoring and for development

potential as a new data source, with high mobile -cellular penetration levels and the increasing

telecommunication data. Indeed, when referring to the data revolution, the United nations High-level Panel cited the example of â€oemobile

technology and other advances to enable real -time monitoring of development results†This section will present some of the existing

data in achieving development goals in various policy areas, including disaster management and sustainable and economic development

Mobile data offer a view of an individual†s behaviour in a low-cost, high-resolution, real

fact that these data are uniquely detailed and tractable, the information captured cannot easily be derived from other sources on such

The fact that the format of the data is relatively similar across different operators and countries creates a huge potential for

potential of mobile data for development in a number of different areas There have been a number of interesting

Mobility data collected immediately after a disaster can in many cases help emergency responders to locate affected populations and

One application of such mobility data is for syndromic surveillance, especially to model the spread of vector-borne22 and

Using mobile data for development A recently published report (Cartesian, 2014) explores the potential of mobile data for development.

It points to three primary types of analysis-ex-post evaluation, real-time measurement, and future predictions and planning-in a

mobile data Migration monitoring Text analysis economic downturn prediction Text analysis commodity fluctuation prediction Assessment

Mobile data to track food assistance delivery Geo-targeted links between Ag suppliers /purchasers Pests, bad

positioning data with malaria prevalence data to identify the source and spread of infections Wesolowski et al.

showed how mobile phone data was used to track the spread of cholera after the 2010 earthquake (Bengtsson et al.

The integration of mobility data from mobile networks with geographic information frameworks, 23 supplemented with additional

A data-centric approach to transportation management is already a reality in many developed economies.

are being fed with sensor data from a multitude of sources such as loop detectors, axle counters

public transport card readers and GPS data derived not only from phones but also from public transport and private vehicles (Amini

data, which, despite its limited spatial accuracy cell ID), has great potential for transportation planning.

data from mobile operator Orange to map out citizens†travel routes in Abidjan, the largest city in

and show how data-driven insights could be used to improve the planning and management of transportation services, thereby

Passive mobile positioning data has also been used for transportation planning and management in Estonia (Ahas and Mark, 2005

How mobile network data can track population displacements †an example from the 2010 Haiti earthquake

was produced on the basis of mobile network data to show the potential of big data in tracking population movements

Leveraging mobile network data for transportation and urban planning in Sri lanka Very similar findings between the results of an official household

results of a big data analysis using mobile-phone data (left-hand map) underscore the merits of big data.

based on mobile-phone data, depicts the relative population density in Colombo city and its surrounding regions at 1300

Mobile big data (left) versus official survey data (right Both passive and active positioning data are used to

analyse traffic conditions, particularly in urban areas with higher base-station density. Active positioning data (especially GPS) produce higher precision in

location data and are therefore the most useful Operators may offer such specialized services based on passive or active location data) either

directly, or by providing data to third parties. Mobile network data are less expensive, are in real time and

are less time-consuming to produce than survey data, particularly in urban and peri-urban areas

where base-station density tends to be high In another example, the analysis of mobility

flows between two Spanish cities derived from three different data sources †mobile -phone data, geolocated Twitter messages and

the census †showed very similar results, and although the representativeness of the Twitter geolocated data was lower than the (real-time

mobile-phone and census data, the degrees of consistency between the population density profiles and mobility patterns detected by

means of the three datasets were significant Lenormand et al, 2014 Chapter 5. The role of big data for ICT monitoring and for development

190 Big data for socioeconomic analysis Data from mobile operators can provide insights in the areas of economic development and

socioeconomic status, often in near real time Big data techniques can therefore complement official statistics in the intervals between official

surveys, which are usually relatively expensive and time-consuming and therefore carried out infrequently. In many cases, insights derived

variables derived from mobile network data to people†s socioeconomic and income levels. The model took into account existing socioeconomic

and income-related data derived from official household surveys, and the results showed that populations with higher socioeconomic levels

model to estimate income levels based on data from mobile network operators Another study, by Gutierrez, Krings and Blondel

data, namely subscriber communication data and airtime credit purchase records, to assess socioeconomic and income levels.

The authors used airtime purchase records based Box 5. 7: Poverty mapping in CÃ'te d†Ivoire using mobile network data

In CÃ'te d†Ivoire, researchers used mobile network data specifically communication patterns, but also airtime credit

purchase records) from Orange to estimate the relative income of individuals, as well as the diversity and inequality of income

regarding revenue data and the difficulty this poses for outside parties wishing to obtain such

data The use of mobile-operator TGD can also foster financial inclusion by facilitating the provision of

relational data typically involve both a limited number of people and limited number of time

patterns based on telecommunication data it has become possible to obtain insights into societal structures on a scale that was

Using mobile-phone data to track the creditworthiness of the unbanked Cignifi, a big data startup, has developed an analytic platform

their mobile-phone data. The Cignifi business model is founded on the idea that â€oemobile phone usage is not random †it is highly

phone calls, text messages, data usage and, extrapolating from these, lifestyles †the company identifies patterns and uses

respond to changes in customer activity as the data are refreshed usually every two weeks. In addition to updating a person†s

findings from the model against historical lending data from approximately 40 000 borrowers using the mobile operator Oi†s

However, telecommunication data are also revolutionizing the study of societal structures at the micro level.

is possible to assess friendship using data from mobile network operators, and that the accuracy is compared high when with self-reported

data. Leveraging these behavioural signatures to obtain an accurate characterization of relationships in the absence of survey data could

also enable the quantification and prediction of macro and micro social-network structures that have thus far been unobservable

There is a case to be made for analysing data captured by telecommunication operators in the interests of improving the current range of

mobile-broadband subscription data is that they do not refer to unique subscriptions, or mobile

Survey-based data, for example on Internet users and mobile-phone users, do not entail the same issues as subscription data.

They are collected through household surveys, directly from citizens, and their level of reliability is relatively

Survey-based data can also be broken down by individual characteristics, including gender, age, educational level and occupation

these data is that they are not widely available in particular, many developing countries do not yet collect data on individual use of

or household access to, ICTS), are relatively expensive to produce, and are much less timely

than subscription data (often with a time lag of one year. Consequently, data on users of

the information society and the types of online service they consume are limited, and in many

administrative data from operators and survey data collected by NSOS, it is particularly interesting to assess some of the ways in

which big data can be used to overcome the shortcomings of existing key ICT indicators and to provide additional insights into ICT access

data to produce new insights. In particular operators†big data could produce information in the following areas

data could provide additional information on gender, socioeconomic status and user location Information on gender or age, for example, could

data techniques applied to users†consumption information, as well as other data contained in customer registration information.

In addition, the analysis of customers†mobility patterns will often allow for an understanding of important locations

All subscription data could provide information as to location. In the case of fixed-telephone and fixed-broadband

surveys for collecting the corresponding data, and the declining response rates where traditional surveys are concerned (Groves

collection of survey data. This could include targeting a wide variety of respondents covering the full spectrum of appropriate

Particularly rich possibilities exist where data from mobile-cellular and mobile-broadband subscriptions are concerned, since they are

voice, data and VAS †over time, and the intensity of use. Mobile operators are able to provide

In addition, mobile-operator data could be combined with customer information from popular online services, such as Facebook

the profiles developed using data from online services with customer profiles generated from analyses of mobile-operator data.

This would require telecommunication operators, OTT providers and other Internet content providers to work together and share information

By applying big data techniques to survey data and administrative data from operators, new insights could be derived, in particular, in respect

of the following Subscriptions versus subscribers: Big data techniques could help extrapolate the actual number of unique mobile subscribers or users

account usage patterns or data from popular Internet companies such as Google or Facebook By linking data collected from different sources

and combining subscription data and usage patterns, a correlation algorithm could be developed to reverse engineer approximate

values for these indicators, in order to estimate user numbers in between surveys, and possibly in real time. This could be pursued in a similar

variables using mobile-phone usage data, as described in greater detail at the beginning of this section. It is important to note here that

data methods only complement existing surveys rather than replacing them completely (see Section 5. 5 for a further discussion of this

information on existing ICT data, and provide new insights into the measurement of the information society.

combining data from surveys with big data to build new correlation and predictive analytic techniques Finally, it should be noted that the methods that

replace survey data, which is needed to build and test correlations and to validate big data results

present what is analytically possible, data access and privacy considerations are complex and nuanced, and therefore place constraints on

exponentially growing data deluge of varying structure and variety comes with its share of challenges.

Data curation, standardization and continuity Data curation and data preparation help to structure, archive, document and preserve

data in a framework that will facilitate human understanding and decision-making. Traditional curation approaches do not scale with big

data and require automation, especially since 85 per cent of big data are estimated to be unstructured (Techamerica Foundation, 2012

Dealing with large heterogeneous data sets calls for algorithms that can understand the data shape while also providing analysts with some

understanding of what the curation is doing to the data (Weber, Palmer and Chao, 2012

Telecom network operators themselves have to contend with interoperability issues arising from the different systems (often from different

vendors) they employ. It is not uncommon for operators to write customized mediation software to overcome potential inter-comparability issues

among data from different systems. The problems are compounded when one has to take account of secondary third-party users that may seek to

leverage the data. The framework used by an NSO to organize data would be different from that

used by a network engineer or a marketing or business intelligence specialist. Naturally, telecom network operators have curated their data

based on their needs. To be able to use telecom big data for development and monitoring and to guarantee its continuity, the creation of

Accessing and storing data, and data philanthropy Big data for development is still in its nascent

access to what is essentially private data. Private corporations would hesitate to share information on their clients and their business processes

the necessary parameters as to how data are to be used, including the manner in which they

challenges associated with extraction of the data on account of different curation approaches and problems relating to the interoperability of

towards sharing data more publicly. Orange for example, hosted a â€oedata for Development Challenge, †releasing an aggregated anonymized

using Orange data from Senegal, is planned for 2015.30 In 2014, Telecom italia initiated a similar challenge, making data from the territories of

Milan and the Autonomous Province of Trento available to researchers for analysis. 31 It has

partnered with other data providers to curate and release additional big datasets containing weather, public and private transport, energy

event and social network data. In both the Orange and Telecom italia cases, researchers had to go through an approval process in order to

regular and safe sharing of data by building on the precedents being created by the ad hoc

the efforts being made to open up private-data stores in order to obtain actionable development insights There is a gap that needs to be addressed if

large-scale pooling and sharing of such data are to become a reality. Cross-sector and cross

and data-curation practices when pooling data from multiple sources. This facilitatory role may even be played by a third-party organization

faced by data providers, effectively acting as a gatekeeper to ensure that data are used transparently and in a way that contributes to

overall scientific knowledge generation, while ensuring that any safeguards that may be applicable in respect of private information are

trial data. To facilitate the process, they hired Yale University†s Open Data Access (YODA) Project

197 Measuring the Information Society Report 2014 to act as gatekeepers (Krumholz, 2014. YODA undertakes the necessary scientific review of any

make use of the data and ensures that necessary privacy and data usage guidelines are followed. 32

The question remains as to who is placed best to act as gatekeepers and standard-bearers when it comes to telecom network big data

third-party data seekers. Telecom operators for their part, are regulated mostly by sector -specific regulators who can also have purview

data-reporting responsibilities of operators Ultimately, however, the decision as to who takes on the gatekeeper and standardization

to play in building an institutional model for data sharing and collaboration, in consultation with all

protocols) of privately held data such as mobile -phone records can be mutually beneficial to both government and private sector.

Depending on the data volume, storage can still be costly, especially where privacy considerations preempt the use of specialized third-party

As social scientists look towards private data sources, privacy and security concerns become paramount. To mitigate the potential risks, all

cases are the primary producers of such data through their activities. It is also the public that

must ultimately decide on how the data they produce may be used. The World Economic Forum†s â€oerethinking Personal Data†project

new data ecosystem. 33 Discussions must address the individual†s privacy expectations, as well as those of private-sector stakeholders looking

identifiable individual (data subject) †(OECD 2013). ) The result of such an approach has been the policy of â€oeinform and consent†practised by

most companies to inform users of what data are Chapter 5. The role of big data for ICT monitoring and for development

data paradigm, the greatest potential often lies in secondary uses, which may well manifest long after the data was collected originally.

It is thus impractical for companies to have a priori knowledge of all the potential uses and to seek

Given the volumes of data that individuals are now generating, companies would find themselves struggling to maintain

privacy issues that may arise when data from one source is combined with data from other sources

to reveal/infer new data and insights. This blurs the lines between personal and nonpersonal

information, allowing seemingly nonpersonal data to be linked to an actual individual (Ohm 2010). ) Digitized behavioural data crumbs may in

fact greatly diminish personal privacy. The use of DPI, for example can technically reveal all of a user†s online activity.

Going one step further it is possible to understand a person†s needs behaviours and preferences by using data

-mining techniques on the digital breadcrumbs For instance, a recent study showed how Facebook â€oelikes†could accurately predict a range

Data anonymization35 (i e. methods designed to strip data of personal information), employed by computational social scientists, has been

called into question (Narayanan and Shmatikov 2008). ) A recent study of mobile CDRS for 1. 5

out that the data could in fact be de-anonymized completely by cross-referencing them with other

data sources. The attendant privacy concerns about such cross-referencing are clear, and have to be taken seriously and addressed

data are kept secure. Data breaches undermine consumer confidence and hinder efforts to exploit big data for the greater social good.

Encryption virtual private networks (VPNS), firewalls, threat monitoring and auditing are some potential technical solutions that are employed currently

approach that shifts governance from the data per se to its use; acknowledges the importance of context rather that treating privacy as a binary

unstructured) data from a multitude of sources â€oemessiness†is to be expected. As Mayer -Schã nberger and Cukier (2013) note, â€oewhat

Data quality and their provenance do matter, and the question is important in establishing the

Data provenance and data cleaning Understanding data provenance involves tracing the pathways taken by data from the

originating source through all the processes that may have mutated, replicated or combined the data that feed into the big data analyses.

This is no simple feat. Nor, given the varied sources of data that are utilized, is it always as feasible as

the scientific community would wish. However at the very least it is important to understand

some aspects of the origin of data. For example the fact that some mobile network operators choose to include the complete routing of a call

that has been forwarded means that there may be multiple records in the CDRS for the same

establish data provenance as envisaged by scientists, it is at the very least important to

have created the data Data cleaning remains a key part of the process to ensure data quality.

It is important to verify that the quantitative and qualitative (i e categorical) variables have been recorded as

expected. In a subsequent step, outliers must be removed, using decision-tree algorithms or other techniques.

However, data cleaning itself is a subjective process (for example, one has to decide which variables to consider) and not a

Are the data representative Related to the question of data provenance is the issue of understanding the underlying

Chapter 5. The role of big data for ICT monitoring and for development 200 population whose behaviour has been captured

The large data sizes may make the sampling rate irrelevant, but they do not necessarily

considering how telecommunication data may be used for monitoring and development While the potential benefits to be gained from

leveraging mobile network operator data for monitoring and development purposes hinges on the large coverage, close to the actual

data does not automatically mean that issues such as measurement bias and methodology internal and external data validity and data

interdependencies can be ignored. These are fundamental issues not just for â€oesmall data†but also for â€oebig data†(Boyd and Crawford, 2012

personas, so studying people†s data exhaust may not always give us insights into real-world dynamics.

where in essence the data artefact is itself a by-product of another activity. Telecom network big data, which mostly fall under this category

of behavioural data are subject to a form of the Heisenberg uncertainty principle, whereby as soon as the basic process of an analysis is

when CDR data from Rwanda showed low mobility in the wake of flooding he theorized that this was due to an outbreak

-data analyses based on big data. For example prior research had established a power-law distribution between the frequency of airtime

with â€oesmall data, †in this case the statistics collected by the Centers for Disease Control and

note, when combined with small data, â€oegreater value can be obtained by combining GFT with other near†real time health data. †Where data

from mobile network operators are used for syndromic surveillance, as in the case of malaria in Kenya (Wesolowski et al.

how telecommunication network data could be used for monitoring, surveys and supplemental datasets will remain important to sharpen the

prior to data anonymization to build a training dataset. This enabled them, for example, to understand variations in mobility, social networks

Virseda (2012) needed census data to build their algorithms and provide training data for their

algorithms to reverse engineer approximate survey maps. Official statistics will thus continue to be important to building the big data models

The fact that the original private data may in many cases not be available to everyone

private-data sources (in a manner that addresses potential privacy concerns) or of peer reviews that can hone

data calls for a combination of specialized skills in the areas of data mining, statistics and

domain expertise, as well as data preparation cleaning and visualization. NSOS may have deep statistical skills in house, but this is not

data sources (UNSC, 2013. Currently, there is a mismatch between the supply of and demand for talented individuals with the requisite

i e. data scientists. Mckinsey predicts that by 2018 the demand for data-savvy managers and analysts in the United states will

amount to 450 000, whereas the supply will fall far short of this, at only 160 000 (Manyika et al

data sources for monitoring the information society As with other official statistics, it is paramount

the applications of data use for development are concerned, operators also have an interest in maximizing the economic well-being of their

share their data with those from other sources including from competitors), but this is something that is worth exploring.

some of their customer data with credit bureaux Governments Governments have different opportunities and different roles to play in the exploitation of big

data for monitoring and development. They can use big data to identify areas where rapid intervention may be necessary,

to encompass best practices in relation to data curation and metadata standards. To this end NSOS must also prioritize the upgrading of the

In return, data users would be permitted to reuse personal data for novel purposes where a privacy assessment indicates

by ensuring that data holders allow others to access their data under fair and reasonable terms

205 Measuring the Information Society Report 2014 International stakeholders International stakeholders †including UN agencies and initiatives (such as ITU and UN

partnerships between data providers and the ICT statistical community, including ITU, could be formed to explore new opportunities and

international data comparability and standards As one of the main international bodies working on issues related to the telecommunication

and engagement on privacy frameworks, data sharing, and analytical global best practices ITU could help reduce the transaction costs

contracts for obtaining data access as well as standards on how the data are stored, collated and curated can collectively reduce the overall

transaction costs of accessing and leveraging telecommunication big data for social good Academia, research institutes and develop

The research into how telecom data may be used to aid broader development is being done mainly by academia, public and private research

with a view to using their data for development They therefore understand the potential and

data for development becomes mainstreamed 207 Measuring the Information Society Report 2014 Chapter 5 Annex

The mobile-telecommunication data that operators possess can be classified into different types, depending on the nature of the

data, service access detail records, location and movement data, device characteristics, customer details and tariff data

Traffic data Operators use a range of metrics to understand and manage the traffic flowing through their

networks. These include •Data volume: both uplink and downlink volumes for Internet traffic can be captured at various levels of

disaggregation down to the individual subscriber, or even to the level of a base station (in the case of a mobile operator

or local switch (in PSTN networks. These can be analysed to understand subscriber demand for data at both an individual

level and at aggregate levels, and the understanding thus gained can be used for billing purposes and for network

Erlang data are used to understand the load on a base station at any given time

The most common use of such data is for basic billing purposes, in addition to which they can

Location and movement data Mobile networks can, depending on their sophistication, capture a range of movement and location variables,

active positioning data (see Annex Box 1 Annex Box 1: Active versus passive positioning data

Passive positioning transaction generated data (TGD) is automatically generated by the network and captured in the

operator†s logs for billing and network management purposes to understand network load and to keep track of the handset in

Active positioning data (which is of relevance only to mobile networks) is location and movement data that is captured in response to a specially initiated network

query to locate a handset using either network or handset-based positioning methods. GPS location data can also be considered

as active positioning data Active positioning data Active positioning data can be generated using either device

-centric or network-centric methods, as well as via satellite i e. GPS. The use of these methods has developed either in

response to national regulations requiring operators to capture higher-precision location data, and/or to a business case for

providing location-based services. The large-scale capture of such higher-resolution data is undertaken mainly by operators

in developed economies. Operators in developing economies use some of these methods, but often on a case-by-case basis

and not for their entire subscriber base. 43 However, this trend is currently changing, and an increasing number of regulators are

Passive positioning data Passive location data are contained in the subscriber†s registration information, which includes some form of mailing address

billing address in the case of postpaid. While fixed-telephone network operators have access only to static location data

mobile networks have much richer and dynamic location data CDRS, SMS detail records and Internet access records are the

main sources of passive positioning data for mobile operators and reside in their data warehouses. These records include the

ID of the antenna (cell ID), which in turn has a geolocation an azimuth (i e. antenna orientation information) and an

angular tilt. 45 It is also possible to obtain such data in real time through data-mediation services,

but these are not universally implemented Passive location data from the billing records are obviously

sparse and generated only when the phone is used and when the network knows which cell a particular handset is currently

these data if they do not have a business case to justify the additional storage costs.

-handoff data provide a time-stamped sequence of cells that the phone was attached to, and provides for a rich mobility profile

Passive positioning data based on cell IDS is inexpensive when compared to active positioning data, but the tradeoff

lies in their lower precision, usually at the level of network cells. This lower-resolution location estimate can range from

such passive location data, at an aggregate level (temporal and/or spatial) these data remain very valuable

Source: ITU Device characteristics All mobile user devices used to access mobile telecommunication services come with an

The Aoa method uses data from base stations that have been augmented with arrays of smart antennas.

Such location data have high spatial resolution, but are costly for operators to implement Enhanced observed time

items of demographic data during the customer registration process. These can include the customer†s age, gender, billing address and

Tariff data Operators maintain the complete tariff sheet and billing records for their current and past services

Mobile operators can associate such data with traffic data to understand the revenue that is being generated by specific network elements

e g. base stations), not only retrospectively, but also, possibly, in real time Chapter 5. The role of big data for ICT monitoring and for development

calls for a data revolution that â€oewould draw on existing and new sources of data to fully integrate statistics into decision-making

and use of, data and ensure increased support for statistical systems. †This suggests that efforts to

and complement, official statistics have turned to the search for new data sources, including big data. In addition, the European Statistical System Committee (ESSC) in 2013 adopted the Scheveningen Memorandum

on the protection of individuals with regard to the processing of personal data and on the free movement of such data

to which personal data cannot legitimately be processed without the consent of the data subject, except if necessary to

of the data subject, who should be given clear and comprehensive information as to the manner and purpose of such

/7 A good example of this is the Conference Board Help Wanted Online (HWOL) data series that measures the number of new

http://www. conference-board. org/data/helpwantedonline. cfm 8 See http://bpp. mit. edu /9 According to Peerreach. com, 20 per cent of Indonesia†s online population uses Twitter, the second highest ratio in the world

15 Deep packet inspection (DPI) is a process that utilizes specialized software to scan all of the data packets traversing a

and analysis of health data about a clinical syndrome that has a significant impact on public health, with the data in question being used to drive decisions about health policy and health education

22 A vector-borne disease is a disease that is transmitted through an agent (person, animal or microorganism

23 A geographic information framework is a representation framework that codes the components of geospatial data in a

standardized manner so as to facilitate analyses and data exchange 24 See http://www. itu. int/en/ITU-D/Statistics/Pages/intlcoop/partnership/default. aspx for more information regarding the

30 More information about the Data for Development (D4d) challenges using Orange data can be found at

32 More information about the Yale university Open Data (YODA) project can be found at http://medicine. yale. edu/core/projects/yodap/index. aspx

42 DPI is a process that utilizes specialized software to scan all of the data packets traversing a particular IP network.

the data deluge makes the scientific method obsolete. Wired Magazine, 16 (7 Askitas, N. and Zimmermann, K. F. 2009), Google econometrics and unemployment forecasting.

Azman, A s.,Urquhart, E. A.,Zaitchik, B. and Lessler, J. 2013), Using Mobile phone Data to Supercharge Epidemic Models of

Data for Development: Net Mobi 2013 Bengtsson, L.,Lu, X.,Thorson, A.,Garfield, R. and von Schreeb, J. 2011), Improved response to disasters and outbreaks by

tracking population movements with mobile phone network data: a post-earthquake geospatial study in Haiti. PLOS

mobility and optimizing public transport using cellphone data. Data for Development: Net Mobi 2013 Blumenstock, J. E. 2011), Using mobile phone data to measure the ties between nations.

Proceedings of the 2011 iconference pp. 195†202. New york, USA: ACM Press. doi: 10.1145/1940761.1940788

Blumenstock, J. E. 2012), Inferring patterns of internal migration from mobile phone call records: evidence from Rwanda

How Does driven Data Decisionmaking Affect Firm Performance? Retrieved from: http://papers. ssrn. com/sol3/papers. cfm?

Byrne, C. 2013), How The UN€ s New Data Lab in Indonesia Uses Twitter to Preempt Disaster, Fast Company.

http://www. fastcolabs. com/3007178/open-company/how-uns-new-data-lab-indonesia-uses-twitter

Caceres, N.,Wideberg, J. P. and Benitez, F. G. 2007), Deriving origin†destination data from a mobile phone network.

Calabrese, F.,Di Lorenzo, G.,Liu, L. and Ratti, C. 2011), Estimating origin-destination flows using mobile phone location data

Global Mobile Data Traffic Forecast Update, 2013-2018, White paper, February 2014 Available at: http://www. cisco. com/c/en/us/solutions/collateral/service-provider/visual-networking-index-vni/white

D. 2009), Inferring friendship network structure by using mobile phone data. Proceedings of the National Academy of Sciences of the United states of america, 106 (36), 15274†8. doi:

using search engine query data. Nature, 457 (7232), 1012†4. doi: 10.1038/nature07634 Groves, R. M. 2011), A possible data future for the observational social sciences.

COSSA 30th Anniversary meeting Washington, D c Gruber, H. and Koutroumpis, P. 2013), Competition enhancing regulation and diffusion of innovation:

How innovative enterprises extract value from uncertain data. Retrieved from: http://public. dhe. ibm. com/common/ssi/ecm/en/gbe03519usen/GBE03519USEN. PDF

ITU (2011), Handbook for the collection of administrative data on Telecommunications/ICT. Available at http://www. itu. int/en/ITU-D/Statistics/Pages/publications/handbook. aspx

Krumholz, H. M. 2014), Give the data to the People, New york times. Retrieved 2 february 2014 from

http://www. nytimes. com/2014/02/03/opinion/give-the-data-to-the-people. html? hp&rref=opinion& r=1

Laney, D. 2001), 3d data management: Controlling data volume, velocity and variety. Gartner, 6 Lazer, D.,Kennedy, R.,King, G. and Vespignani, A. 2014), Big data.

The parable of Google Flu: traps in big data analysis. Science New york, N y.),343 (6176), 1203†5. doi:

Data for Development: Net Mobi 2013 Mayer-Schã nberger, V. and Cukier, K. 2013), Big data:

Naef, E.,Muelbert, P.,Raza, S.,Frederick, R.,Kendall, J. and Gupta, N. 2014), Using mobile data for development.

Pindolia, D. K.,Garcia, A j.,Wesolowski, A.,Smith, D. L.,Buckee, C. O.,Noor, A m.,Tatem, A j. 2012), Human movement data

Using New Data to Understand Emerging Vulnerability in Real-time UN High-level Panel (2013), A new global Partnership:

A. 2013), Undefined By Data: A Survey of Big data Definitions, 2. Databases. Retrieved from http://arxiv. org/abs/1309.5821

Weber, N m.,Palmer, C. L. and Chao, T. C. 2012), Current Trends and Future Directions in Data Curation Research and Education

Using Cellular network Data Data for Development: Net Mobi 2013 Xia, J. 2011), The third-generation-mobile (3g) policy and deployment in China:

Current status, challenges, and prospects Telecommunications Policy, Volume 35, Issue 1 february 2011, pp. 51-63. doi:

objectives, data availability and the results of various statistical analyses such as the principal component analysis (PCA.

Data for all of these indicators are collected by ITU. 2 1. Fixed-telephone subscriptions per 100

data cards or USB modems, subscriptions to public mobile data services, private trunked mobile radio, telepoint, radio paging and

telemetry services 3. International Internet bandwidth (bit/s) per Internet user International Internet bandwidth refers to the

Data are obtained by countries through national household surveys and are provided either directly to ITU by national statistical offices

There are certain data limits to this indicator insofar as estimates have to be calculated for many developing countries which do not yet

more data become available, the quality of the indicator will improve 5. Percentage of households with Internet

Data are obtained by countries through national household surveys and are provided either directly to ITU by national statistical offices

There are certain data limits to this indicator insofar as estimates have to be calculated for many developing countries which do not yet

more data become available, the quality of the indicator will improve b) ICT use indicators

Data for all of these indicators are collected by ITU. 4 1. Percentage of individuals using the Internet

Data are obtained by countries through national household surveys and are provided either directly to ITU by national statistical offices

There are certain data limits to this indicator insofar as estimates have to be calculated for many developing countries which do not yet

more data become available, the quality of the indicator will improve 2. Fixed (wired)- broadband subscriptions per

mobile-broadband data subscriptions to the public Internet. It covers actual subscribers, not potential subscribers

with advertised data speeds of 256 kbit/s or greater that allow access to the

been used to set up an Internet data connection using Internet Protocol (IP) in the past three months.

Internet data connection, even if the messages are delivered via IP. Dedicated mobile-broadband data subscriptions refers to subscriptions to dedicated data

services (over a mobile network) that allow access to the greater Internet and which are purchased separately from

voice services, either as a standalone service (e g. using a data card such as a USB modem/dongle) or as an add-on data

package to voice services which requires an additional subscription. All dedicated mobile-broadband subscriptions with

recurring subscription fees are included regardless of actual use. Prepaid mobile -broadband plans require use if there is no

Data on adult literacy rates and gross secondary and tertiary enrolment ratios are collected by the

data A critical step in the construction of the index is to create a complete data set, without missing values

There are several imputation techniques that can be applied to estimate missing data. 6 Each of the

imputation techniques, like any other method employed in the process, has its own strengths and

ensure that the imputed data will reflect a country†s actual level of ICT access, usage and skills

missing data, where previous year data are not available to calculate the growth rates. Hot -deck imputation uses data from countries with

â€oesimilar†characteristics, such as GNI per capita and geographic location. For example, missing data for country A were estimated for a certain indicator by

first identifying the countries that have similar levels of GNI per capita and that are from the same region

use data of country A was estimated by using Internet use data of country B from the same region

with similar level of GNI per capita and similar level of fixed Internet and wireless-broadband

estimate missing data for all indicators included in the index 3. Normalization of data Normalization of the data is necessary before any

aggregation can be made in order to ensure that the data set uses the same unit of measurement For the indicators selected for the construction

of the IDI, it is important to transform the values to the same unit of measurement, since

some values are expressed as a percentage of the population/of households, whereby the maximum value is 100, while other indicators

of the huge dispersion of values, the data were transformed first to a logarithmic log) scale. Outliers were identified then

After normalizing the data, the individual series were rescaled all to identical ranges, from 1 to

the data were transformed first to a logarithmic (log) scale The ideal value of 787†260 bit/s per Internet user is equivalent to 5. 90 if transformed to a log scale

weighting and aggregation of the data Each of the processes or combination of processes affects the IDI value.

1 Principal component analysis was used to examine the underlying nature of the data. A more detailed description of the

2 More information about the indicators is available in the ITU †Handbook for the collection of administrative data on

EGH. pdf). As some of the data used in the calculation of the IDI were collected before that meeting,

however, the data may not necessarily reflect these revisions 4 More information about the indicators is available in the ITU â€oehandbook for the collection of administrative data on

telecommunications/ICT€ †2011, see ITU 2011b and the ITU â€oemanual for Measuring ICT Access and Use by Households and

Annex 2. ICT price data methodology 1. Price data collection and sources The price data presented in this report were

collected in the fourth quarter of 2013. The data were collected through the ITU ICT Price Basket

questionnaire, which was sent to the administrations and statistical contacts of all 193 ITU Member States in October 2013.

Through the questionnaire contacts were requested to provide 2013 data for fixed-telephone, mobile-cellular, fixed-broadband

and mobile-broadband prices; the 2011 and 2012 prices were included for reference, where available. For those countries that did not reply

The collection of price data from ITU Member States and the methodology applied for the IPB was agreed upon by the ITU Expert

Annex 2. ICT price data methodology 232 The fixed-telephone sub-basket does not take

6. The same price plan should be used for collecting all the data specified. For example, if a given Plan A is selected for the fixed

for data collection, since it may not be possible to isolate the prices for one service. It is preferable to use prices for a specific

Annex 2. ICT price data methodology 234 Annex Table 2. 1: OECD mobile-cellular low-user call distribution (2009 methodology

data usage of (a minimum of) 1 Gigabyte GB). ) For plans that limit the monthly amount

of data transferred by including data volume caps below 1 GB, the cost for the additional

Annex 2. ICT price data methodology 236 Annex Box 2. 3: Rules applied in collecting fixed-broadband Internet prices

7. The same price plan should be used for collecting all the data specified. For example, if a given Plan A is selected for the fixed

the volume of data that can be downloaded, etc 8. Prices should be collected for regular (non-promotional) plan

for price data collection, since it may not be possible to isolate the prices for one service.

GB of data volume. If providers set a limit of less than 1 GB on the amount of data that can be

transferred within a month, then the price per additional byte is added to the monthly price

so as to calculate the cost of 1 GB of data per month. Preference should be given to the most

-broadband price data from ITU Member States was agreed upon by the ITU Expert Group on

lessons learned from the first data collection exercise. The revised methodology was endorsed by the eleventh World Telecommunication/ICT

To capture the price of different data packages covering prepaid and postpaid services, and supported by different devices (handset and

collected for two different data thresholds, based on a set of rules (see Annex Box 2. 4

of data allowance (or validity. The customer i) continues to use the service and pays an

offers that include the minimum amount of data for each respective mobile-broadband plan The guiding idea is to base each plan on what

data allowance and validity of each respective plan Annex 2. ICT price data methodology 238

Annex Box 2. 4: Rules applied in collecting mobile-broadband prices11 1. Prices should be collected based on one of the following technologies:

plans, the plan satisfying the indicated data volume requirement should be used 8. Where operators propose different commitment periods for postpaid mobile-broadband plans,

9. Price data should be collected for the cheapest plan with a data volume allowance of a minimum of

Data volumes should refer to both upload and download data volumes. If prices are linked to †hours of use†and not to data volumes, this information should be added in a separate note

Note: ITU will most likely not be able to include these cases in a comparison

10. A validity period of 30 days should be chosen. If this is not available, 15 days should be used.

Preference should be given to packages (including a certain data volume. Pay-as-you-go offers should be used

since most often there are limits in the data volumes, either applied by throttling (limiting the speed)

of ITU supply-side indicators (i e. data collected from operators), as well as to discuss outstanding methodological issues and

EGTI is open to all ITU members and experts in the field of ICT statistics and data collection.

8 Data for fixed-telephone, mobile-cellular and fixed-broadband have been collected since 2008 through the ITU ICT Price

10 Some operators throttle speeds after the data allowance included in the base package has been reached.

Data in italics refer to ITU estimates. For further notes, see p. 248 Source: ITU World Telecommunication/ICT Indicators database

Data in italics refer to ITU estimates. For further notes, see p. 248 Source: ITU World Telecommunication/ICT Indicators database

Data in italics refer to ITU estimates Source: ITU World Telecommunication/ICT Indicators database Annex 3. Statistical tables of indicators used to compute de IDI

change in tariff policy of the biggest WLL operator. 13) Data excluding own (NRA) consumption. 14) Excl. voice-over-IP (Voip

on 2013q3 data. 19) Refers to active Fixed Wired/Wireless lines. 20) Per June 2013.21) Operators†data. 22) Residential:

Definitive data (annual report) may change because quarterly reports use a smaller sample of operators than annual report. 24) Fixed and fixed-wireless subscriptions. 25) Excl. internal lines and WLR of

Data for the third quarter of 2013 Mobile-cellular subscriptions per 100 inhabitants, 2012 1) Numbers are down due to data cleanse. 2) ACMA Communications Report 2011-12.2) incl. payphone, excl.

Voip. 3) Active subscriptions. 4) Bhutan Telecom and Tashi Cell are the only two service providers in Bhutan. 5) Activity criteria:

voice or data communication in the last month. 6 december 2012.7) Total number of subscriptions (including non-active:

Excl. 2 720 698 prepaid cards that are used to provide Travel SIM/World Mobile service. 9) Excl. data-only SIM CARDS and M2m

Including PHS and data cards undividable. 15) Decrease was due to registration of SIMS. 16) Figure obtained from all five mobile (GSM & CDMA) operators

Incl. inactive. 22) Incl. data-only subscriptions not possible to disaggregate the information at this point.

active and non-active subscriptions. 28) Incl. data dedicated subscriptions. 29) Decrease due to the closing of MTS

but excl. mobile data subscriptions (via data cards, USB modems and M2m cards). 7) Estimate

8) Incl. data-only subscriptions. 9) Data based on NRA estimates. Excl. data-only SIM CARDS and M2m cards. 10) Incl. fixed

wireless local loop (WLL) subscriptions. 11) Incl. PHS and data cards, undividable. 12) Active subscriptions (after clearing

simcards that became inactive. 13) Figure obtained from all four mobile (GSM & CDMA) operators currently providing

16 february 2013 NTA MIS. 17) Based on 2013q3 data. 18) Measured using subscriptions active in the last 90 days. 19) Per

Definitive data (annual report) may change because quarterly reports use a smaller sample of operators than annual report. 26) Data for the fourth quarter of 2013.27) Incl. data dedicated subscriptions

28) Reduction is due to implementation of sim cards registration International Internet bandwidth Bit/s per Internet user, 2012

MOT. 11) Data obtained from nine service operators. 12 may 2012 purchased capacity. Lit capacity: 43 096 471 Mbit/s. 13) Incoming capacity;

weekly incoming capacity, averaged over 4 weeks in December 14) Preliminary. 15) SLT Data. 16) Refers to the total capacity

4 may be revised with comprehensive data from mobile broadbnad providers. 5) Estimate. 6 june. 7) Sum of incoming capacity of all ISPS in the country. 8) Activated external

capacity. 9) By September 2013.10) Data obtained from eight service operators. 11) 1st april 2013 purchased capacity.

or laptop. 4) Data correspond to dwellings (not households). 5) Ghana Living Standards Survey 2012/2013

-Economic survey-2012.9) Census data. 10) Computer includes the number of personal computer, Notebook, and PDA

1) Labour force Survey 2013.2) Cambodia Inter-censal Population Survey. 3) Refers to PC, laptop or a tablet. 4) Data

%2) Preliminary. 3) Data correspond to dwellings (not households. 4) Ghana Living Standards Survey 2012/2013.

via mobile modem. 10) Census data. 11) Excl. households which didn†t know type of internet access 172 346 households

1) Labour force Survey 2013.2) Corresponds to all type of internet connections. 3) Data correspond to dwellings (not

instead of data survey. 13) Ghana Living Standards Survey 2012/2013. The estimate is based on weighting households who use

based on the data provided by 89.1%of operators 6) The figure is corrected. The previous figure was 1†636†700.7) Only ADSL, excl. cable modem. 8) Speeds greater than, or

than 144 kbit/s. 14) Operators data/ictqatar estimate. 15) Incl. subscriptions at downstream speeds equal to, or greater than

subscriptions. 18) Excl. corporate connections. 19) Data reflect subscriptions with associated transfer rates exceeding 200

and Fiber links. 5) Estimate, no specific data collected for â ¥ 256 kbit/s. 6) CRC estimation as of 31.12.2013.7) Estimate. 8) Data

Broadband in India with effect from 18th july 2013.12) December 13) ADSL and Leased lines. 14) Based on 2013q3 data

15) Per June 2013.16) Operators†data. 17) Incl. subscriptions at downstream speeds equal to, or greater than, 144 kbit/s (the

Definitive data (annual report may change because quarterly reports use a smaller sample of operators than annual report. 19) Estimate.

2013.20) Excl. 3175 Wimax subscriptions. 21) Excl. corporate connections. 22) 2013 data is an estimate as of June 30, 2013

Data reflect subscriptions with associated transfer rates exceeding 200 kbps in at least one direction, consistent with the

12) Satellite, BWA and active mobile subscriptions. 13) Estimate based on partial SIT data and ITU estimates. 14) Speeds

/Data refer to the sum of fixed wireless broadband and active mobile -broadband subscriptions. 16) Incl. mobile broadband and Wimax. 17) Estimate. 18) ETL and LTC. 19) Incl. narrowband

but data are not available. 23) Operators data/ictqatar estimate. 24) Refers to active mobile-broadband

effect from 18th july 2013.16) Data refer to the sum of fixed wireless broadband and active mobile-broadband subscriptions

17) 2013 data is an estimate as of June 30, 2013.18) Incl. mobile broadband and Wimax. 19) Estimate based on 1. Standard

mobile subscriptions using data services 2. Dedicated data subscriptions 3. Add on data packages. 20) Based on 2013q3

data. 21) Per June 2013.22) Mobile broadband only. Fixed wireless and satellite exist but data are not available. 23) Operatorsâ€

data. 24) As at Dec 2013.25) Q4 report. Definitive data (annual report) may change because quarterly reports use a smaller

sample of operators than annual report. 26) Wireless Broadband services are not being offered in St vincent as yet.

We anticipate that Mobile broadband and terrestrial fixed broadband services would be in place by the end of 2014.27) OFCOM

5. 3 Telecommunication data and their potential for big data analytics 5. 4 Big data from mobile telecommunications for development and for better monitoring

Annex 2. ICT price data methodology Annex 3. Statistical tables of indicators used to compute the IDI

< Back - Next >

Overtext Web Module V3.0 Alpha
Copyright Semantic-Knowledge, 1994-2011