How to Use Data to Solve Social Problems and Drive Decisions: 4 Takeaways from Data for Good Exchange 2016

These days, the normal reaction I get after telling someone that I work at John Snow, Inc. is not “Oh right, the broad street pump guy!” I get Game of Thrones references instead. While the crowd at Bloomberg Data for Good Exchange 2016 (#D4GX) may be GOT fans, they are, more importantly, data scientists; connoisseurs of collecting, analyzing, and using data to inform and change systems…like the father of epidemiology himself.

#D4GX was just that; a conglomeration of folks wearing public and private sector hats, coming together to share best practices, new innovations, and opportunities for collaboration in their respective industries, from urban planning, and fishing and forestry, to public health.

So, how can we use data to address social issues?

1. Development convergence.

Have you seen the SDGs? They’re no joke, and 2030 will be here before you know it. It’s time to combine forces and share data. Development convergence encourages industries to come together to synthesize their data, ideas, methods, and dollars. Sounds like #D4GX don’t you think?

Panel, Decoding the SDGs: Data Solutions for Achieving the Sustainable Development Goals. Timothy A. A. Stiles (KPMG), Robert Kirkpatrick (UN Global Pulse), Marcia Odell (Plan International), and James Baker (Clinton Climate Initiative).

• For example, the UN Global Pulse has launched the Data for Climate Action campaign where “companies commit to share data in an act of data philanthropy” to fight climate change. This act of convergence will increase industries’ ability to leverage anonymized big data in hopes of mitigating climate change (SDG 13).
• On that note, KPMG and Plan International have partnered to keep gender inequity on the radar and better monitor progress on gender targets around the world. “The initiative will utilise existing and new quantitative and qualitative data, and monitor strategically chosen gender-related SDG indicators to track the progress being made for girls and women across key lifecycle stages,” according to Plan International.

2. Maps are so in right now.

It’s been a couple hundred years since Dr. Snow plotted cholera deaths on a simple map; one of the earliest examples of data visualization and epidemiology. But using maps, geo-spatial data and geographic information system (GIS) data is trendier than ever. Think: satellites, drones, smart phones. These powerful technologies offer billions of pixels worth of data that many industries have only just tapped into as a resource. The high number of public beta mode presentations demonstrated that there is an immense amount of potential to be unlocked in this arena. For example:

• The University of Michigan needed information on city infrastructure in order to assess the risk of lead contamination of service lines for individual homes in Flint—turns out these were in the form of annotated maps. Speaking of infrastructure, Stanford is teaching computers how to read satellite data to predict poverty based on housing material. Think: tin roofs vs. straw.

Flint City Service Line Records, University of Michigan.

• On a public health note, this is cool. The Real Impact Analytics group is plotting telecom mobility data on maps to track the movement of phone users to keep an eye on the incidence and spread of malaria in Zambia in order to better predict out breaks or re-introduction of the disease.
  

  Source: Real Impact Analytics.

3. More management.

With big data comes big responsibility. It is up to data scientists to manage the knowledge at hand to drive decision making. Data needs to be pushed out and shared to contribute to society for the greater good.

• That means if you’ve developed an intricate GIS tool to monitor fishing activity in global waters, you take action when your data tells you that a ship is illegally fishing in a protected area. It also means that you build the capacity of local government to use this data to track their waters and enforce their own policies.
• Our own presentation highlighted IMPACT Team Networks, an innovative intervention that builds a network of better leadership and management of public health supply chain data by local managers at the county level. We also showed how better data visibility and management can serve as a mechanism to lower the cost of public health supply chains and improve performance indicators like stockouts in Tanzania.
  

  JSI’s Cary Spisak presents on IMPACT Team Networks in Kenya as a part of Quantified Communities Paper Presentations.

4. Data ethics.

Technology and data are powerful tools that when analyzed and applied can inform and inspire change for good. However, data, and its overarching algorithms, can also be considered a weapon of math destruction (as coined by keynote speaker and mathematician Cathy O’Neil). It is crucial that data scientists remove biases or wishful thinking from their analyses so that findings are honest and pure. O’Neil suggested that data scientists “need a Hippocratic oath for modeling” to do no harm. This was tied in the development convergence conversation; how do we ensure data anonymity and ethical use of big data once it becomes increasingly available?

• This is nothing new for the medical or public health data scene; key step in data set inception is to protect patient privacy and anonymize. Some examples from other industries include the transportation data gained from commuter cards (mapping itinerary patterns…could be a little big brother, eh?) and the protracted and controversial growth model score algorithm used to measure teaching ability in the public school system.

Get more information on our #D4GX submissions below:

• Better Data Visibility & Data Use Result in Lower Cost and Improved Performance in Medicine Supply Chains
• A Case in Kenya: Unlocking Bottlenecks in Public Health Supply Chains through Data Dashboards and Enhanced Governance Structures

Learn about the conference.