The coronavirus pandemic disrupted the lives and livelihoods of many people in 2020. COVID-19 impacted everything from education, food supply, healthcare to the job market and housing. In our story, we look at how pressure on renters during the pandemic resulted in a frenzy of concern about homelessness and the disproportionate effect on vulnerable populations. We also look at how layoffs contributed to an already unstable housing situation for many.
This repository shares the data about layoffs and closures in California from 2019-2020. Layoffs and closures have always been a consistent presence in the market place. For the purposes of our story, we analyzed the 2019 data and compared it to 2020 data to assess the difference in job losses from a non-pandemic year to a pandemic year.
In our story, we've focused on a few counties with high rent burdens, where a significant percentage of households pay more than 50% of their income in rent. We also looked into some metropolitan areas that may not have the highest rent burdens, but do have high population density. There are many patterns present in this data we have yet to explore and we are sharing it with the hopes that other data journalists and news organizations will report similar stories of their own - and share their how counties and cities have been impacted by layoffs and closures due to the coronavirus pandemic.
This data was used as a section in a larger evictions story written by Big Local News.
The data was collected from:
To use the code in this project, install the pandas and jupyter/jupyter lab libraries.
git clone git@github.com:biglocalnews/WARN_Cali_analysis.git
pip install jupyterlab pandas
processed_data.ipynb and rent_burden_calculation.ipynb can be run simultaneously. The processed directory holds all WARN layoff records that were downloaded from the Employment Development Department. processed_data.ipynb creates a single file of all those records and exports it back to the processed directory.
rent_burden_calculation.ipynb takes ACS rent as a % of income data and the list of public housing authorities in California. With these two datasets, the rent burden in each county is calculated and merged with he public housing data, creating a table that shows the rent burden for the counties of each housing authority. This file is exported to the rent_burden directory as a full version with all of the columns and a truncated version with only the necessary columns.
Run the following notebooks in order:
data_cleaning.ipynb cleans the data. The process of data-cleaning is outlined at the top of the notebook and throughout. Once the data has been transformed by the notebook, it is ready written to the open_refine directory as clean_warn_data.csv and is ready to be further standardized in OpenRefine.
After you run data_cleaning.ipynb, there is a manual break in the process. The clean_warn_data.csv file has to be imported into OpenRefine and cleaned with the various algorithms made available through the software. open_refine_dedupe.ipynb details the process of cleaning in OpenRefine.
Once the data has been thoroughly processed in OpenRefine, the data then must be exported from OpenRefine and re-added to the open_refine directory of this repository as open_refine_exported_warn_data.csv.
The open_refine_dedupe.ipynb notebook will remove possible duplicate records from the data. The deduplicated data will then be exported to the analysis directory as finalized_warn_data.csv where it will be ready for analysis.
layoffs_analysis.ipynb is where the analysis happens. This notebook grabs the finalized_warn_data.csv and analyzes it to produce the numbers and used in our story.