Takeaways From Teaching Data Journalism

This spring I taught a data journalism class for Hack University, a “non-code school”, to get students familiar with different uses and tools around data journalism. While having a full-time job and volunteering, planning a course with two three-hour sessions a week for eight weeks was a lot of work. 

I’ve never taught a class before. Even though my class was only eight students, it was a challenge. I learned a lot being in front of a class, but these are my two biggest takeaways.

Talk Less

Going into the class I was extremely worried how I was going to keep the attention of the class and keep the discussion going. I sought out some advice MacGregor Campbell. He gave me a piece of advice that I would pass on to anyone teaching for the first time: “talk less.”

When leading discussion it was better for me to briefly talk about a data project or visualization and then ask a few questions. If I had bullet points to explain, the class would naturally bring them up simply by asking good questions. This made the teaching/learning process more organic and engaging. It seemed whenever I started to “lecture,” the class lost interest.

Do More

There were a few classes where I wanted to show different tools or concepts and then lead a discussion. But, it was much more successful to think of an exercise that would cover the concepts and the students would learn the concepts by doing.

The challenge is thinking of exercises that actually teach the right concepts and are engaging. There were two exercises that I thought worked well:

  1. Use two Google forms that ask the same questions but with different types of responses to show the pro’s and con’s of data collection. See form one here and two here.
  2. After a class on data visualization, I gave everyone in the class the same data set (in this case it is was historic Oregon voter turnout), and asked them to make a chart that best visualized the data. Then the class presented and critiqued. As a followup, they merged the data with another data set that they thought would provide more context to voter data.