As legislators try to combat the deleterious effects of gerrymandering and redlining, studies must be done thereafter to measure whether the areas demographics or class distribution actually improve.
Using the most updated census data collected in 2020, previously redlined zones can be studied to see whether their composition or "desirability" to live in has changed since the redlining started in the 1940s. The goal of this project is to see whether the dissolution of the redlined districts impacted the communities' demographics after.
If successful, this classifier will be able to inform as to whether previously redlined districts need further social programs or help. If it finds that these districts haven't changed much in their demographics, it's implied that the socioeconomic situations of those in these areas have not changed or improved.
A negative outcome of this classifier could be that social initiatives for social and economic equality could be deterred by the evidence showing that the abolition of older, racist policies won't fix things anyways, and, as such, will deter the motivation for social change.
More literature on the lasting effects of redlining as well as current, commonplace efforts to deinstitutionalize this anti-black rhetoric can be found here:
The dataset used was compiled by FiveThirtyEight and consists of the Home Owners' Loan Corporation's redlined zones and their respective grade of desirability i.e. how much one would want to live there, as well as updated census demographic data from the latest 2020 U.S. Census. It is linked here.
import pandas as pd
url = 'https://raw.githubusercontent.com/fivethirtyeight/data/master/redlining/metro-grades.csv'
df = pd.read_csv(url)
df.head(5)
| metro_area | holc_grade | white_pop | black_pop | hisp_pop | asian_pop | other_pop | total_pop | pct_white | pct_black | ... | surr_area_white_pop | surr_area_black_pop | surr_area_hisp_pop | surr_area_asian_pop | surr_area_other_pop | surr_area_pct_white | surr_area_pct_black | surr_area_pct_hisp | surr_area_pct_asian | surr_area_pct_other | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | Akron, OH | A | 24702 | 8624 | 956 | 688 | 1993 | 36963 | 66.83 | 23.33 | ... | 304399 | 70692 | 11037 | 17295 | 23839 | 71.24 | 16.55 | 2.58 | 4.05 | 5.58 |
| 1 | Akron, OH | B | 41531 | 16499 | 2208 | 3367 | 4211 | 67816 | 61.24 | 24.33 | ... | 304399 | 70692 | 11037 | 17295 | 23839 | 71.24 | 16.55 | 2.58 | 4.05 | 5.58 |
| 2 | Akron, OH | C | 73105 | 22847 | 3149 | 6291 | 7302 | 112694 | 64.87 | 20.27 | ... | 304399 | 70692 | 11037 | 17295 | 23839 | 71.24 | 16.55 | 2.58 | 4.05 | 5.58 |
| 3 | Akron, OH | D | 6179 | 6921 | 567 | 455 | 1022 | 15144 | 40.80 | 45.70 | ... | 304399 | 70692 | 11037 | 17295 | 23839 | 71.24 | 16.55 | 2.58 | 4.05 | 5.58 |
| 4 | Albany-Schenectady-Troy, NY | A | 16989 | 1818 | 1317 | 1998 | 1182 | 23303 | 72.91 | 7.80 | ... | 387016 | 68371 | 42699 | 41112 | 40596 | 66.75 | 11.79 | 7.36 | 7.09 | 7.00 |
5 rows × 28 columns
from collections import defaultdict
data_attributes = defaultdict()
data_attributes['metro_area'] = 'official name of zone'
data_attributes['holc_grade'] = 'the Home Owners Loan Corporations grade for each zone, assigned in the 1940s. It is now considered to be a racist form of arbitrary districting.'
data_attributes['white_pop'] = 'estimated number of white people living within the zone'
data_attributes['black_pop'] = 'estimated number of black people living within the zone'
data_attributes['lq_white'] = 'location quotient for a specific demographic and zone which indicates how over- or underrepresented this ethnicity is compared to surrounding communities'
data_attributes
defaultdict(None,
{'metro_area': 'official name of zone',
'holc_grade': 'the Home Owners Loan Corporations grade for each zone, assigned in the 1940s. It is now considered to be a racist form of arbitrary districting.',
'white_pop': 'estimated number of white people living within the zone',
'black_pop': 'estimated number of black people living within the zone',
'lq_white': 'location quotient for a specific demographic and zone which indicates how over- or underrepresented this ethnicity is compared to surrounding communities'})
As seen above, many of the features present detail the ethnic makeup of a zone as well as information on how said ethnic makeup compares to surrounding districts. It should be noted that the x_pop and lq_x attributes are present for each of the five ethnic categories (white, black, Hispanic, Asian, and other), but I've only added a few for succinctness.
With these data, I think it would be possible to try to evaluate whether the zones' redlining had a lasting effect on the population demographics even after redlining practices stopped, as the demographic data of the locations and surrounding locations would well inform whether the zones' makeups had changed.
We propose that the data in this set could be analyzed using a regression method. This way, changes in the zones' demographics can be approximated into a line of best fit in order to elucidate whether the change in a zone was notable or not.