How Big Data can Strengthen Official Statistics in Africa: A view from Pulse Lab Kampala
The use of digital devices whenever we go about our daily lives is increasing at an unprecedented pace in Africa. The explosion of mobile phone ownership for example is leading to an exponential growth of Internet usage on the continent. Research estimates that in Africa, mobile data use will increase twenty times from what we currently see by the end of the decade, twice as much as the anticipated global expansion.
As a result of the increased use of digital devices, a wide range of new data sources are available to support official statistics. To make the deluge of new data sources and the methodologies derived from them useful for development, requires a new mind-set of approaching official statistics.
The UN Statistical Commission agreed at its 45th meeting (in March 2014) to create the Global Working Group (GWG) on Big Data for Official Statistics. In its first meeting, the GWG acknowledged that: ‘using Big Data for official statistics is an obligation for the statistical community based on the fundamental principle to meet the expectation of society for enhanced products and improved and more efficient ways of working’.
So how do we change mind-sets and apply Big Data for official statistics? The key is to look at the potential of new data sources to complement traditional stats with real-time, targeted updates. This year, at Pulse Lab Kampala, we aim to illustrate how Big Data can be used to generate statistics to monitor the Global Goals. The Lab has begun working on two tools that showcase the synergy of big data and traditional statistics, which can be used by national statistical bureaus.
Automated Identification and Counting
In Northern Uganda, where poverty levels are high and a majority of the population lives in rural areas, a proxy indicator of poverty identified by the Uganda Bureau of Statistics is roofing material. Traditional thatched roofs harbor pests and diseases and require a high degree of maintenance. As the household economy improves, families often upgrade their dwelling by changing from the traditional grass thatch to iron sheets.
Pulse Lab Kampala has developed a prototype to measure poverty trends with “automated roof counting” through satellite imagery. The Lab has developed an image processing software that identifies roofs and the type of material they are made of. Initial partners on this project have included the University of Edinburgh, UNDP, the Uganda Bureau of Statistics, the National Planning Authority of Uganda, as well as the support of several private sector companies. This tool can be used alongside surveys taken by the Uganda Bureau of Statistics to provide new insights in real time.
Changes in forest area, land degradation or land desertification can be observed with the human eye looking at images. If the human eye can identify a change, trained software can measure it over time. Pulse Lab Kampala is working on a tool that uses satellite imagery to identify climate change effects, with a special focus on the extent of wetlands. The tool will quantify changes of various indicators over time. Partners on this pilot include UN Agencies in Uganda, Makerere University, and the Uganda Bureau of Statistics. For this project, the Lab will build on its previous work building a user-friendly interface for leveraging spatial data to monitor biodiversity in Zimbabwe (http://www.unglobalpulse.org/blog/new-big-data-tool-help-policymakers-pr...).
Pulse Lab Kampala aims to support monitoring and achievement of the Global Goals with data innovation. Strengthening official national statistics by tapping into new types of data will ensure better tracking mechanisms and contribute to better lives.