19  Storytelling

Settling in

• Sit with your project groups.
• Put away laptops and phones until the activity.

Goals

• Work toward Project Milestone 3: progress report.
• Do some project exploratory data analysis (EDA), i.e. preliminary data checks, cleaning, and analysis.
• Start thinking about your work in the context of telling a story with data. This is the overarching project goal!

19.1 Activity

There is no recipe, and are no specific directions, for telling a story with data. There are lots of ways to tell a good story…and lots of ways to tell a bad story. This fivethirtyeight.com article provides one decent example. The format isn’t exactly like that of your final report (eg: it doesn’t include code). But it includes a lot of the big pieces.

1. Start the story

Name the 2 important things that the authors provide in paragraph 1.

2. Discuss your data

• What data did the authors use in their analysis?
• What info did the authors include about the data?
• Where is this information in the article?

3. Start simple. Provide preliminary insights.

Starting simple is important for you and the audience!

• What question(s) does this plot answer?
• What does it accomplish?
• What follow-up questions does it prompt?

4. Specify what you want the reader to learn!!!

YOU are telling the data story. It’s your job, not the readers’ job, to make meaning from your analysis. Showing but not discussing a plot is pointless. Provide clear takeaways.

5. Zoom in / consider different angles / address follow-up questions

What new angle did the authors consider in the follow-up plot? Where did they zoom in?

6. Provide numerical summaries for more detail

7. Zoom in some more / iterate

8. Conclude & provide clear takeaways.

Communicate what key point(s) you want the readers to have learned from your analysis.

19.2 Next steps

• Chip away at project milestone 3. Review it on Moodle first!

• As you know already, it’s not easy to collaborate on a qmd! If you continue on with Data Science (or Statistics or Computer Science), you’ll want to eventually learn the GitHub collaboration / version control software. Some basics are included here.

• Orient your work around the idea of telling a story with data. This requires exploration and iteration:

  • start with something simple
  • after each new numerical or graphical summary, discuss with your group: what does it answer? what new questions does it prompt? where should we zoom in next?
  • after each new graphical summary, ask whether some numerical summaries would provide important details
  • think about what you want the audience to take away from your analysis

• Discussion

  • I’m going to sit down with each group for ~5 minutes. During that time, I won’t be bouncing around to help with troubleshooting. Work as a group to troubleshoot.
  • After meeting with every group, I’ll then bounce around based on where there are questions. But I’ll only address code specifics if the entire group has already tried. Working together through roadblocks, identifying helpful resources, etc is part of the project.

19.3 Wrap-up

• Due Tuesday by 1:10pm (10 minutes before class): Project Milestone 3

19.4 Solutions

Click for Solutions

1. Start the story

motivation and research question

2. Discuss your data

Data is discussed in paragraph 3, not paragraph 1. The authors share where they got the data, the time span of the data, and roughly how much data they got.

3. Start simple

• Q: What was the overall change in pickups for each car category?

• This gives us a sense of scale and overall prompts.

• Lots of follow-up questions. eg: How did these changes differ by neighborhood?

5. Zoom in

How did the cab/uber changes differ by neighborhood?