This week the Microfinance Information Exchange (MIX) launched southafrica.mixmarket.org mapping over 40,000 financial points of service all the way down to the sub-place (the equivalent of town), continuing their effort to provide transparency and analysis to subnational financial landscapes in the developing world. This site specifically focuses on South Africa and aims to present the most comprehensive view of financial providers for the country.
Much like previous collaborations for Nigeria and Africa as whole, the site shows not only where these services are available, but also where they are in relation to other relevant socio-economic indicators such as population and poverty. This is a classic example where mapping the data helps show complex comparisons across indicators. Of special interest is the data from Finscope, which includes both the percentage of adults who do not use any formal or informal financial services and the percentage of adults who have never banked.
In MIX’s blog post on the site, Scott Gaul explains:
The Finmark demand surveys can give us more detail about the use of financial services: is there demand for financial services in those regions? Again, perhaps not surprisingly, the regions that are the least well-served are also those with the fewest branches. But we can also segment the data by provider - for instance, Postbank branches are already located in the regions where the most people have never been banked (i.e. there is a positive correlation).
Over 40,000 points of service are represented on the map, encompassing all types of credit providers. To obtain this unprecedented amount of financial data for the region MIX employed a web scraper that collected publicly available addresses from several primary sources. Much of the remaining time was then dedicated to preparing them for mapping. The entire dataset is made available on the site in the form of a fusion table and CSV file.
The cumbersome nature of the source data presented us with several issues that only served to highlight the need for this data to be open and in an easily digestible format. There remain roughly 5,000 locations that failed to make it on the map due to poor formatting or missing data. Two estimations - locations per 100,000 adults and locations per 1,000 square kilometer - that were derived from this dataset must be considered incomplete as a result.
How we built it
These tools will aid MIX as they continue to bring the financial landscape of more developing countries into focus. The biggest remaining hurdle is collecting and processing data that is diverse and at times hard to reach. We hope the success of this site compels more organizations to make their data more accessible.