Capstone

TO DO

• Implement geographic similarity thing
  • sample function header:

    • # ID = the id of county of block
      # type = {'county' or 'block'}
      # colName = {'%hispanic' or ...}
      get_avg(int ID, str type, str colName){
          // compute average of target variable for specified region
          return theAverage
      }
      

  • some ideas:
    • just group by block
    • just group by county
    • group by county but assign a higher weight to regions in the same block
    • group by X but assign a higher weight to regions in the same block/County that have similar diverity scores
      • note that the diversity score has not yet been created
• Diversity score
  • given a list of variables that indicate diversity and the id of the region we are trying to quantify how diverse it is...
    • compute national average for the given variables
      • note that doing this read/computation every single time will take awhile, consider computing the averages and storing them in a file or dictionary
    • compute Eucledian distance between target region and national average
    • This value will probably still need to be normalized at the end
• Misc
  • implement include_strings in load_data()
  • Add y intercept to model
  • normalize all of the values
    • NOTE the percentage columns are normalized on the range [0, 100]. Need to be consistent