This repository contains data and code to reproduce the findings featured in our story, "L.A.’s Scoring System for Subsidized Housing Gives Black and Latino People Experiencing Homelessness Lower Priority Scores."
Our methodology is described in "How We Investigated L.A.’s Homelessness Scoring System."
We cleaned the original CRPA response from the Los Angeles Homeless Services Authority using notebooks/0-clean.ipynb to create a cleaned file called data/assessments.csv, which we used for our analyses. Outputs for our analyses, which were used for graphics in our story, are available in the output folder.
Jupyter notebooks used for data preparation and analysis are in the notebooks folder.
This notebook takes the original CPRA response, which we did not include in this repo due to privacy concerns, and outputs a cleaned version called data/assessments.csv, which we use in our analyses.
This notebook takes data/assessments.csv and performs several analyses, including chi-square tests for both the Next Step Tool and the CES Survey Part 1. It outputs several .csv files we use for graphics in both our story and methodology, all of which are available in the output folder.
This notebook takes data/assessments.csv and performs linear and logistic regressions for both the Next Step Tool and the CES Survey Part 1. For more details, please refer to our methodology.
This notebook takes data/assessments.csv and calculates two sets of subscores: one using all assessmesnts, and another using only assessments without discrepancies between our calculated score and the actual score in the dataset. For more details, please refer to our methodology.
This notebook takes data/assessments.csv and performs everything done in notebooks/1-analysis.ipynb, but instead of using TOTAL_SCORE (including to calculate Acuity), it uses CALC_TOTAL_SCORE. It outputs several .csv files, all of which are available in the output/calc_total_score_note folder.
All notebooks have already been run.
We used Python 3.9. A list of libraries we used, and which versions they are, is available in requirements.txt. If you already have Python 3 installed, you can install these packages by running pip install -r requirements.txt.
You can then re-run our notebooks by running nbexec notebooks/1-analysis.ipynb notebooks/2-regressions.ipynb.
A data dictionary for data/assessments.csv can be found in docs/assessments-dict.md.
While we do not include the original CPRA response we recieved from LAHSA due to privacy concerns, we include a data dictionary at docs/cpra-response-dict.md.