Data for Development: the D4D Challenge on Mobile Phone Data

Abstract

The availability of detailed mobility traces and mobile phone communication data for large populations has already had a significant impact on research in behavioral science. Some researchers consider such datasets as an opportunity to refine the analysis of human behavior [5], while others question the usefulness of such datasets to draw conclusions on collective human behavior [1,2,5].Digital traces left by mobile phone users often reveal sensitive private individual information. It is therefore natural to limit access to such data. Limited access to data of scientific interest is however a potential source of a "new digital divide" in the scientific community, as described in [2]. In order to improve the availability of large mobile phone datasets and to foster research in this area, the Orange Group decided to provide anonymized datasets from Ivory Coast for the purpose of scientific research. With around five million customers, Orange has a significant market share in Ivory Coast, whose total population is estimated to be 20 million individuals. In addition to the scientific benefit, the project intends to foster development in Ivory Coast by establishing new collaborations with African scientists and by providing behavioral data that has not yet been collected by the national statistics agency [3].2. Other datasets on Ivory Coast. Researchers participating in the D4D challenge are encouraged to combine the D4D mobile phone datasets with other datasets and source of information. These sources include the following.African Development Bank Group. The African Development Bank (AfDB) Groups mission is to help reduce poverty, improve living conditions for Africans and mobilize resources for the continents economic and social development. With this objective in mind, the institution aims at assisting African countries in their efforts to achieve sustainable economic development and social progress. http://www.afdb.org/en/ African Economic Outlook. Economic, social and political developments of African countries, with the expertise of the African Development Bank, the OECD Development Centre, the United Nations Economic Commission for Africa, the United Nations Development Programme and a network of African think tanks and research centres. http://www.africaneconomicoutlook.org/en/ Africa and Middle East Telecom News. Africa and Middle East Telecom-Week' tracks the fixed, broadband and mobile phone markets in Africa and Middle East. http://www.africantelecomsnews.com/ Africa Renewal on Line. The Africa Renewal magazine is produced by the United Nations organism and provides up-to-date information and analysis of the major economic and development challenges facing Africa today. It works with the media in Africa and beyond to promote the work of the United Nations, Africa and the international community to bring peace and development to Africa.

Patterns, Entropy, and Predictability of Human Mobility and Life

“…The user panel, with its particular socio-economic and geographic background (for a comparison, consider Ref. [33]) and its size may introduce various bias, which however pales in importance if put into contrast with a few bigger issues in the context of human life analysis. We already find noticeable variations in the simplest personal quantitative characteristics: the number of locations to describe the personal life, the number of patterns, and the entropy, in the case at hand.…”

Universal predictability of mobility patterns in cities

“…(3) Abidjan mobile phone users. The data set contains 607,167 mobile phone users' movements between 381 cell phone antennas in Abidjan, the biggest city of Ivory Coast, during a two-week observation period [43]. Each movement record contains the coordinates (longitude and latitude) of the origin and destination.…”

A unified modularity by encoding the similarity attraction feature into the null model

“…In this section, we show community detection results by optimizing Dist-Modularity in three representative social networks: the Zachary's karate club network [53], the bottlenose dolphin network [35,36], and the antenna-to-antenna network of D4D mobile phone datasets [6].…”

“…The third network in our experiments, namely the antenna-to-antenna network, is the one that has additional information on nodes. This network is based on anonymous records of mobile phone calls between five million of Orange's customers ¶ in Ivory Coast between Dec 1, 2011 and Apr 28, 2012 [6]. The nodes represent 1216 antennas which are associated with geographical position information, namely the coordinates of the two-dimensional space.…”