Child pages
  • Computational thinking and social studies

Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

  • Contemporary American political polling
    • Start with a recent dataset from a single pollster, such as YouGov or Ipsos. Study the data structure
      • In what ways is this dataset an abstraction?
      • How has the pollster decomposed the problem?
      • What algorithm or rule is being used to generate the top-line result? 
      • What claim or generalization is being made?
    • Next, move to an aggregator – https://projects.fivethirtyeight.com/trump-approval-ratings seems to have a very clear, accessible layout. 
      • Understand the new, more complex dataset (weights, adjustments, etc.)
      • Look for patterns across pollsters (LV vs. RV vs. A screens, for example)
      • Generate rules or predictions – what will things look like 2 weeks from now? Why? 
    • To turn to a meta-discussion of computational thinking: The aggregator is an abstraction of a set of abstractions. Does this make it more accurate / informative / reliable? If so, why? How does that work? 


  • For computer programming classes: Google and Bing now recognize geo-coordinates as a data type. For example: "40 N, 75 W" produces a map as its first search result – it's just across the river from Philadelphia. This means you can write Python scripts that automate one or more geo-searches. Some possibilities:
    • Given a list of class birthdays, generate a set of maps that show everyone's "birthday location". Example: Someone born on 12/12 would have a birthday location in northeastern Nigeria (12 N, 12 E)
    • Given coordinates for a location (say, a student's home address), generate a map of the lat/lon on the opposite side of the world. (For example, San Diego's lat/lon is approximately 33 N, 117 W. The opposite lat/lon (33 S, 117 E) is approximately Perth, Australia). You can decompose this task by first getting the mirror lat (33 S, 117 W) and/or mirror lon (33 N, 117 E). You can add complexity by indexing the resulting lat-lon to a table of world-wide cities and automatically generating the closest major city. 
    • If you can write scripts for Google Earth, see if you can do the same thing for Mars or the Moon. (For example, the Opportunity Rover is at about 2 S, 5.5 W...and it's not moving from there.)