This script computes the conditional entropy of the main-race partition (X) given transit‐quality bins (Y) for the specified area.
-
Inputs:
- Census file (
race_normalized_<area>.csv):
Columns:zip_code,population, and one column per race proportion. - Transit scores file (
transit_scores_<area>.csv):
Columns:zip_codeand the chosen transit metric (e.g.walk_score). - main_race_thresh: a float in interval of [0, 1]. Zip code block has a race's population proportion greater or equal to the threshold will use this race as its main-race label.
- good_thresh: a float in interval of [0, 1]. Greater or equal to this threshold will be marked as good for transit quality.
- bad_thresh: a float in interval of [0, 1]. Lower than this threshold will be marked as bad for transit quality.
- Census file (
-
Entropy calculation:
- Race partition (X):
- Assign each ZIP to its main race if that race’s proportion ≥
main_race_thresh; otherwise label itmixed.
- Assign each ZIP to its main race if that race’s proportion ≥
- Transit partition (Y):
- Bin each ZIP’s score into
good(≥good_thresh),average(≥bad_thresh), orbad(<bad_thresh).
- Bin each ZIP’s score into
pop_XY[y][x]= total population in ZIPs where X = x and Y = y.pop_Y[y]= total population in ZIPs where Y = y.- For each Y = y:
- Ent(X|Y=y) = sum over x of pop_XY[y][x] * log2(1 / pop_XY[y][x])
- Overall:
- Ent(X|Y) = sum over y of pop_Y[y] * Ent(X|Y=y)
- We can safely do this because all data is normalized to its ZIP code.
- Race partition (X):
-
Output:
- output will be write into (
main_race_results_<area>_<transit_score>.txt) - Prints total population, Ent(X|Y=y) for each bin, full breakdown of contributions,
- Lists ZIP codes in each X and Y partition for debugging.
- output will be write into (
This script will visualize parititions of transit quality.
- Input:
'zip_results_{area}_{transit_score}_{main_race_thresh}.txt'this file will be generated inmain_race.py.- area: String "city" or "county"
- transit_score: String such as "walk_score" or "all_transit"
- transit_score: float between 0 to 1.
- Output:
- It will generate a matplotlib interative graph.
- It will also generate an html file in format of
stl_{area}_{transit_score}_map_{main_race_thresh}.html.
I will briefly introduce the programs included in this folder.
A ZCTA (ZIP Code Tabulation Area) is the Census Bureau’s best‐fit area for a five-digit ZIP Code. You can think of it as a polygon drawn around the majority of addresses sharing the same ZIP. Unlike USPS ZIPs (which are routes, not areas), ZCTAs are static geographic units the Census uses to tabulate and publish population or demographic data at ZIP-like level.
Since all ZCTAs are valid USPS ZIPs, and they are more standardized, we will use ZCTAs as units of the data we collect for this project.
We fetched the St. Louis county and city's ZCTAs from Dexter — Data EXTractER. You may access the data with following link:
Note: Throughout this document, “ZIP code” and “ZCTA” are used interchangeably; however, all references should be understood as ZCTAs.
-
Transit Scores
All transit-related metrics have been min–max rescaled to the [0, 1] interval. -
Census Data
Ethnicity counts were normalized by ZCTA population.
Mathematically, for each ZCTA and each group (e.g., White), we compute:NormalizedValue_{ZCTA, group} = Count_{ZCTA, group} / Population_{ZCTA}
Source of Info:
This script will convert a list of zip codes to the cooresponidng ethinicity and race information from census data. Following is the categories we include:
- American Indian and Alaska Native
- Asian
- Black or African American
- Hispanic or Latino
- Native Hawaiian and Other Pacific Islander
- Not Hispanic or Latino
- Some Other Race
- Two or More Races
- White
Input: a list of zip codes Output: a CSV file called census_race_ethinicity.csv I manually renamed the files and saved each for St. Louis City and St. Louis County.
- For St. Louis City: census_race_ethinicity_city.csv
- For St. Louis County: census_race_ethinicity_county.csv
Source of Info:
- https://www.walkscore.com/
- Python library: pgeocode.Nominatim
This script can directly convert a list of ZIP codes to the corresponding walk and bike scores.
Input: an array of ZIP codes
Output: a CSV file named zip_walk_bike_scores.csv
I manually renamed the files and saved each for St. Louis City and St. Louis County.
- For St. Louis City: zip_walk_bike_scores_city.csv
- For St. Louis County: zip_walk_bike_scores_county.csv
Source of Info:
This script allows you to convert an input array of ZIP codes to the corresponding latitude and longitude coordinates.
Input: an array of ZIP codes
Output: a CSV file called zip_code_to_coords.csv
The file stored in the folder right now is just for testing purposes.
Source of Info:
This script has two functions:
-
Allows the user to input several lines, so they can enter as many latitude and longitude pairs as they want.
- Input: floats in the terminal
- Output: a CSV file called walk_bike_scores.csv
-
Allows the user to input an array of latitude and longitude coordinates.
- Input: an array of coordinates
- Output: a CSV file called walk_bike_scores.csv
The file stored in the folder right now is just for testing purposes.
Source of Info:
This script allows the user to input a pair of latitude and longitude coordinates and retrieve the corresponding walk and bike scores.
Input: floats in the terminal
Output: corresponding walk and bike scores shown in the terminal
Source of Info:
This script allows the user to input a list of zip codes and recrieve the cooresponding AllTransit score.
Input: a list of zip codes Output: a CSV file called allTransit.csv I manually renamed the files and saved each for St. Louis City and St. Louis County.
- For St. Louis City: allTransit_city.csv
- For St. Louis County: allTransit_county.csv