Subscribe to DSC Newsletter

How to analyze spatial data stored in a non-traditional format.

Hello DSC folks,

I have some survey data that I'm trying to use for mapping, but I can't figure out a good way to actually map responses without losing a lot of data I would like to preserve.

We have about 700 student participants and we have the location of where they're currently studying, and then we ask them what regions of the country they would be willing to move to (9 subregions which are 3-9 states each). Each region has it's own column, and the responses are dummy-coded 1 if they checked that answer, 0 if not (they could check more than one). I also have broken those out into 50 columns so each state has its own, with the same responses (so if they checked Middle Atlantic, there would be a 1 in NY, PA, and NJ).

My question is somewhat twofold:

  1. How do I best visualize these data on a map or series of maps? I have some experience with ArcGIS, R spatial, and Tableau but am willing to learn others. 
  2. How do I best model the relationship between current location (or location of institution) and places where they are willing to move after graduation? Is spatial or aspatial analysis better, and how much would I benefit from learning a bit of ML here? 

Any advice would be much appreciated!

Tags: GIS, R, Tableau, data, learning, machine, map, mapping, maps, spatial

Views: 197

Reply to This

Replies to This Discussion

Hi Stephen,

Sounds like the most fine-grained location is state-level, and you mentioned that you already have experience with Tableau, so that would be a good place to start to visualize the data. Maybe start with a nice migration map. Some examples:

Alternatively, a circular plot could show the flows between states nicely. Here's an example:

As far as modeling the relationship between current location and willingness to move, I suppose that would depend on the goals of the study. A few basic ideas: whether the student is willing to leave the state, whether they are wiling to move to an adjacent state (use an adjacency matrix there), the distance they are willing to move (use a 51x51 distance matrix for the states + DC, more if you sampled from PR, GU, VI, etc.), maybe try willingness to move between "salt water" and "fresh water" states, "red state" vs. "blue state" moves, whatever else supports the goals of the study.

Good luck!

RSS

Videos

  • Add Videos
  • View All

© 2019   Data Science Central ®   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service