Sometimes we think of data as something that just exists, that we discover or access. But data sets are shaped by the people who create them, who structure them, and who use them. In order to produce visualizations and analysis that are accurate, critical, and transparent about data sources and bias, we have to evaluate our data sets.

Choose a data set from the list below and answer the following questions in about 300-400 words total. The first four questions are about the sources of the data. You can answer these relatively briefly. If you can’t find the answers to any of them, note that in your review. The last questions are about how we can or should use it, challenges with the data, and any perceptible biases in it. These will take more reflection and should accordingly make up the bulk of your review. Submit the review as a blog post on the class website by Wednesday, March 19 at 1:30, with the tag “Data review.”

Data set profile:

  1. Who is credited as the creator and/or contributors of this data set? Who are they?
  2. What are the sources of their data?
  3. Why did they create or compile it?
  4. How has it been used?
  5. What format is the data set in?

Data set evaluation:

  1. Take a look at the data itself. How have they structured it? What fields have they chosen? What effect might that have on how it can be used?
  2. Read the creators’ description of the data set. Have they described the choices they made in cleaning the data, and if so, how? What effect might those choices have on the data?
  3. Consider the creators’ identities and goals in creating the data set. How might those things have shaped the data, either intentionally or inadvertently?
  4. What would you use this data for?

Data sets to choose from:

Average Weekly Hours of All Employees, STL Metro Statistical Area

Monthly and Seasonal Temperatures, St. Louis (since 1874)

Washington Park Cemetery North Reinterment Index

Tombstone Transcription Project (choose a local cemetery that’s been transcribed)

Illinois Landcover in the Early 1800s

If you have another data set you’d like to review instead, please check with me (the earlier, the better).