In Jake Porway’s recent article “Five principles for applying data science for social good”, he examines the swell in organizations wanting to use “data for good.” Even with a little bit more research, it’s challenging to estimate exactly how many organizations with this mantra. Having at least two events on my calendar with this theme this week, I can say a lot of people are trying to use data for good.
Porway states that ‘Many efforts will not only fall short of lasting impact – they will make no change at all.’ I would agree with him since data and technology by itself without these principles is just data.
Principle 1 is that statistics is more than percentages. Affecting social change requires that we tell the story of how data is more than spreadsheets to non-technical audiences. To effectively tell great data stories, Tableau suggests thinking of the analysis as a story, being authentic, being visual, make it easy for the audience and yourself and invite and direct discussion.
Principle 2 is that finding problems is harder than finding solutions. A barrier to using data for social change is that we don’t have a clear understanding what is possible from data. I would agree with this principle as I recently explored the problem of the high cost of gluten-free foods in my October 19th article. Solving societal problems involves creative thinking and connecting datasets from multiple sources.
Principle 3 is that communication is more important than technology. I would disagree slightly with Porway in saying that communication is just as important as technology. When we gather communities of social change experts and data wrangling experts, there will be plenty of opportunity for misunderstanding. We need to be able to articulate the technical for any audience and willing to speak up if we don’t understand the technical. As Danny Brown has noted, people give context to the data.
Principle 4 is that we need diverse viewpoints. In order to make a lasting social change, you need to understand the problem in its entirety. Unlike in the business world, metrics for social change are not easily defined. Groups like DataKind Labs try to bring together subject matter experts in the societal problem, the data owners and data scientists to create meaningful solutions.
Principle 5 is that we must design for people. I strongly agree with Porway that if we don’t design for people using the data, the work is in vain. We need to take more time to identify which question(s) we’re trying to answer with the data and if we’re even defining the problem correctly. You simply can’t just hack your way to long-term, lasting social change.