Data Science for Democratic Freedoms
Ahead of Data Fest 2018, Erin Akred, discusses how data science can be used to effect positive change in the world today.
While there have been tremendous advances in technology and more data than ever now available, for-profit companies tend to be the only entities with the necessary budget and staff to take advantage. What if the same algorithms that companies use to boost profits or recommend products could instead be applied to some of the toughest challenges the world faces today?
Whether it’s predicting and preventing human rights violations alongside Amnesty International, helping public utilities in California forecast water demand and improve allocation efforts during drought periods, or increasing government transparency by detecting edited content on federal websites, DataKind aims to tackle complex social issues, teaming up expert data scientists with mission-driven organisations on projects that leverage machine learning and artificial intelligence (AI).
But what does it take to make such collaborations successful?
We’ve learned that applying data science for good requires not only relevant data sets, but also relevant decision makers, technical and issue area experts, funders and advocates that can inform and help co-design solutions that will have an impact.
It ‘takes a village’ to leverage such advanced technology for good. In fact, we like to think of it as an ecosystem or what we call the Six Components for Successful Data Science for Good Projects.
Having datasets and a team of data scientists ready to donate their skills for good is of course necessary to achieve this work, however you also need a strong project partner – be it a non-profit, social enterprise or government agency – to collaborate with. These ‘social actors’ as we call them are DataKind’s vehicles for change – they will ultimately be the ones to implement whatever model or tool is developed or use the insights uncovered to inform organisational efforts.
In addition to these social actors, you also need to engage subject matter experts to help advise on the scope of work and shape the problem statement. This formulaic statement is the tool we use to clarify how the scope of work for a project will lead to positive impact the world, specifying what data will be used, what deliverables will be created and how the work will be incorporated into an organisation’s operations.
Last but not least, funding is needed to support the work. We depend on our network of corporate and foundation partners to make this work possible, not only through their financial support but to provide data as well as guidance and expertise in specific issue areas.
Connect the Dots
DataKind is the convener that connects these dots, bringing all these usually far-flung resources and people together.
For example, following the U.S. elections in November 2016, the Environmental Data Governance Initiative (EDGI) began as a small group of concerned researchers working to save valuable information and data that was being deleted from federal websites by the current administration. It has since grown into a network of 116 people from 30 institutions and now monitors tens of thousands of federal environmental agency web pages for changes being made to data available as well as information and how it’s presented.
DataKind is now working with EDGI to automate the classification of these changes – many of which of subtle but significant – , as they could have sweeping and long-lasting effects on federal environmental governance overall.
At DataFest 18, I’ll share learnings and insights from DataKind and the more than 200 projects applying machine learning and AI for the greater good that we’ve completed. I’ll also discuss the types of projects that are better suited toward positively impacting the human condition.
I’ll walk through how each of these six components took shape in our current project with EDGI and what we’ve learned to date about how data science can be applied to protect democratic freedoms in the U.S.
Erin Akred, Data Science Manager at global data science consulting organisation DataKind