PULSE LAB DIARIES

Exploring insights from new data sources to support the SDGs

Paula Hidalgo Sanchis
Oct 12, 2017

In this article, Paula Hidalgo-Sanchis, Manager of Pulse Lab Kampala, highlights a few of the new data exploration projects the Lab has of late embarked on together with UN partners.

Big data refers to the astonishing amount of information that is created as a by-product of the growing digitization of our lives – our use of mobile phones, social networks, search engines, online payment methods, apps, and so on. Many examples exist demonstrating the value of new data sources to target interventions based on real-time information as a source of insights into human behaviour.

In Uganda for example, the Communications Commission estimates that for every landline there are 18 mobile phone connections. The Internet World Statistics estimated 12 million people had access to the Internet in 2016, accounting for around 30% of the population. Facebook users alone reached 150 million across the continent with some 2.2 million in Uganda.

Pulse Lab Kampala works to understand how insights from new data sources can help monitor and achieve the Global Goals in Uganda, and largely on the African continent. Why is exploration important? The challenge of using new digital data sources for sustainable development is that there is no precedent in how to use them, how to analyse them or how to apply them to inform development practice. Innovating is about introducing new methods, ideas or products. If they are new, that means they were not discovered before. In order to discover something new, we need to explore.  

Exploration is key to innovation. During this process, we sometimes unveil a new digital data source that proves to be useful while other times realize that it may be too biased to be used or presents an incomplete analysis. Most times, during the exploration process, we write software that is later used with similar data sources. Bottom line: exploration is key to innovation.

Below are some examples of recent explorations we have been working on at the Lab.

Exploring uses of water and sewerage data

With our partners at the National Water and Sewerage Corporation (NWSC) we analysed consumer complaints recorded through the company’s call centre, to understand trends in these complaints. Our hypothesis was that better understanding of consumer complaints could support better service delivery.

Complaints data containing 500,000 records and 8 fields in a time series of 6 years were analysed. A series of data visualizations were developed to show: the 10 top complaints by all consumers, the difference in complaints by NWSC branches over time, and the response time to address complaints by NWSC branches.

With this exploratory work, the NWSC has deepened their understanding of the type of data that is collected through the consumer complaints centre and its usability.  The results of this analysis have not been made public due to privacy considerations.

Exploring data generated by anti-robbery GPS devices

GPS devices installed in boda bodas (local motorcycle taxis) as anti-robbery devices generate vast amounts of data as they share geo-referenced data every 15 seconds including longitude, latitude and time. This data is collected over the course of several days for 24 hours every day.

The data can inform on the conditions of roads – by using the maximum speed of any boda-boda on a road section. For example, an analysis conducted at the Lab showed that road sections with higher maximum speeds appear to be in better condition than those with lower speeds.

Data derived from GPS devices can also be an indicator of road congestion as we can monitor the average speed of all boda-bodas for a particular road section. Our analysis showed that higher average speeds indicated less congested roads, while lower average speeds signalled sections where traffic jams occur. Human mobility is relevant to development in many ways and understanding mobility patterns can help in the definition of urban planning strategies.

Exploring digital trade data

The Lab collaborated with the Artificial Intelligence Research Group (AIR) on developing a mobile auction market that links farmers to traders, called Kudu.

Through the app, any farmer or trader anywhere in the country can send a message in the system using any type of phone. The application computationally matches compatible buyers and sellers and informs them of who they have been matched with. By matching farmers and traders with the use of algorithms, challenges such as the inefficient flow of information, or dependency on intermediaries can be overcome. The system encourages competition in the market and limits unnecessary movement of both farmers and traders.

At the Lab, we used the data generated from real transactions on agricultural purchases to develop a visualization that shows the volume of trade flows between districts in Uganda. The data set analysed consisted of 11,729 records, and contained transactions from April 2016 to October 2016.

Image shows district trade flows for October 2017.