- 8th August, 2017
Doing Good With Data
We live in an age where data collection and analysis are increasingly central to our lives. We rely on data applications to tell us the fastest route to work, the movie we’re most likely to enjoy on Netflix and which books to avoid on Amazon. But how are applied data driven approaches improving public safety or increasing access to critical services such as healthcare or clean drinking water?
Doing good is hard. Applying analytics to societal endeavours brings with it several unexpected hurdles, some of which are described by Jake Porway of Datakind. Datakind is a US based non-profit organisation that uses data analytics to address societal problems by bringing together government (who are often the custodians of the data), social entities (who are typically intimately familiar with the issues) and data scientists (who have the analytic skills). These collaborations are critical for bridging skill and knowledge gaps as well as for introducing diverse points of view.
However, even when such collaborations are facilitated, articulating the problems which could be addressed with the available data remains challenging. Social or governmental stakeholders may be unclear about how their data could be used, and therefore the challenges it could inform. Similarly, despite technical knowhow and the best of intentions, data scientists won’t necessarily be well acquainted with the complexities associated with tackling humanitarian problems.
Many shades of grey. The problem is that social issues are complex and interrelated. According to Porway, tackling social issues is like trying to unravel a knotted ball of string: you can’t pull on one thread without tightening up the remaining knot. Furthermore, goals are often subjectively good or bad rather than absolutely right or wrong and so end goals are often poorly defined. It’s also worth remembering how different a non-profit’s business environment is from the for-profit business model. Very often these organisations are already stretched to their limits and cannot afford to do more than they are already, let alone hire additional data analytics resources.
Communication is another hurdle. Data scientists can become focussed (read obsessed) with the technical aspects of the architecture or their pet predictive model and unfortunately, cryptic jargon can dominate our discourse. The net effect can be to intimidate others with socially orientated areas of expertise into mystified silence.
Data, data everywhere. Where socially orientated analytics are being practiced, they appear to be almost exclusively under the umbrella of state funded academia. That said, there doesn’t appear to be a shortage of open data. In South Africa for example, state funded organisations like Stats SA, the Human Sciences Research Council and others offer many interesting datasets, often for free and at appropriate granularity. openAFRICA is doing a great job of compiling data from varied resources and making them accessible. Companies like CodeforAfrica and OpenUp are spearheading projects using open data to inform decision making processes. One very interesting example is the living wage calculator, which can give employers better insights into fair living wages for domestic workers.
So how can data scientists working in for-profit organisations contribute to social upliftment? One approach might be to engage in more philanthropic pursuits. Forbes recently discussed some of the benefits organisations can realise from such initiatives.
At Ixio Analytics, we collaborate closely with resource centers like Insight2Impact to help drive financial inclusion in developing countries. Furthermore, we’re working to disseminate data science instruction more widely across Africa by partnering with corporates and institutions of higher learning. While data science is unlikely to be humantity’s solve-all elixir, it certainly has the potential to be an indispensable tool in the policy maker’s and NGO’s toolbox.