5 takeaways:

➀ With Census data, precision counts. Be mindful about whether you’re measuring income or healthcare access for a neighborhood in a city or an entire state. Boundaries matter, Sean McMinn told NPF Widening the Pipeline fellows. “You don’t necessarily have to have a vaccination site on your exact block to have access to healthcare. But if there’s nothing in your neighborhood, that’s a little bit more concerning and probably becomes more of a hassle or a challenge to get that healthcare.” Know the difference between a Census block, tract, and block groups, and remember that the smaller you get in assessing a geographic area, the higher margin of error those Census statistics have. [Transcript | Video]

➁ Decide what you’re trying to communicate with the story. If it’s about people vulnerable to extreme heat, you’d want heat data from satellites, McMinn said. If you’re trying to find out who has access to vaccines, ask for data on vaccination sites.  The data needs to be geographic, which means there should be a latitude or longitude, or some type of proscribed area associated with it. And McMinn said the information should come to you in the form of “shapefiles,” which are data sets pertaining to things like race, income, etc.

➂ The next step is QGIS—and there’s a hack. McMinn urged Widening fellows to explore the Quantum Geographic Information System, the free, open-source cross platform geographic information system that supports viewing, editing, printing and analysis of geospatial data. It can be intimidating, McMinn acknowledged, but “You just have to ignore the vast majority of the buttons and worry about learning a couple of buttons that do what you want it to do.”

➃ Dig a bit deeper, and patterns emerge. Say you have a list of addresses for COVID vaccination sites in your city. The way to get inside that geographic data is called geocoding, McMinn said. And whether you know it or not, you’re utilizing it each time you rely on Google Maps or Yelp.  “Geocoding uses a piece of software that’s already developed by someone else to turn addresses into latitude and longitudes.”

➄ Stick with it. Using the tools to decipher demographic data may be challenging, but McMinn shared a piece of advice he got during his Paul Miller Fellowship year, from veteran data journalist and University of Maryland lecturer Derek Willis.  He said, “The only difference between someone who learns code or data analysis or mapping and the person who never does is that when not if, but when you get stuck, one person keeps going and the other person gives up.”  

The Widening the Pipeline fellowship is sponsored by the Evelyn Y. Davis Foundation, Bayer AG, J&J and Twitter. NPF is solely responsible for the content.