diego domenzain September 2020 @ Colorado School of Mines

Given a data set in n-dimensions, how do we visualize it in our limited human vision?

These scripts are examples of self-organizing maps.

• synth_ex.m is a light-weight synthetic example.

• zodiac_ex.m uses data from one of those buzz-feed-type questionnaires but for a work environment.

  • Columns are the questions, rows are people.
  • Interesting trends:
    • Persons 1, 9, and 14 were the oldest ladies in the group.
    • Persons 2, 3, and 8 were officemates (all dudes).
    • Persons 5, 6, 7, and 11 were the most boring in the group by far (ladies and dudes).
    • Persons 10, 12, and 13 were international students (all dudes).

• geoclasses_ex.m uses data from geoscience students (not necessarily geo-physicists).

  • Columns are classes they take, and rows are attributes from those classes.
  • Interesting trends
    • Classes 2 and 12 are hydrology and geophysics.
    • Classes 1, 3, 8, and 14 are sed/strat, geomorphology, mineralogy and earth materials.
    • Classes 4 and 6 are structure and field.
    • Classes 9 and 10 are historical and paleontology.
    • Classes 11 and 13 are environment and climate.