These are highlights of the code I wrote while a graduate student at the Howard Center for Investigative Journalism. I did not know any programming languages prior to starting my graduate degree in Sept. 2019. I am now proficient in R and Python, and also have experience visualizing the data findings with Adobe After Effects and Datawrapper.
I served as the lead data journalist on our fall project looking into evictions by public housing authorities. The analysis was done using a combination of court records we scraped ourselves and internal records we requested from the public housing authorities.
The main finding was that nonpayment of rent was the leading reason tenants were evicted from public housing. The housing authorities were among the leading eviction filers in the counties we looked at, some of them filing against the same tenants several times each year.
The final code shared here was written collaboratively, but I did the work to get the team through the analysis.
Read the story.
Take a look at the findings.
During our summer project, we looked into evictions that were violating the CARES Act moratorium during the COVID-19 pandemic. For this project, I learned python in order to build webscrapers to gather court data from 2019 through 2020 in the select counties we were investigating.
Once the scrapers ran, I analyzed the court records for the top filers and any landlords who had filed across state lines. I also pulled in the weekly census pulse surveys to look at the demographics of who was beign impacted by these violations. Through that, I found more people of color missed rent payments in June than white people.
Read the story.
My first project with the Howard Center, I was the data analyst for Washington, D.C. Using Open Data D.C., I pulled down 311 data going back to 2010 to analyze homeless-related calls. What I first noticed was that the specific service code for homeless calls had been eliminated in June 2015. After that, I had to use a string of search terms in the "details" section of each call to try to identify which ones were related to homeless encampments or individuals.
I also geocoded the latitude and longitude of the calls to be able to use the census data to get demographics about where the calls were coming from. Through that, I found the calls mostly came from areas with higher median incomes.
Read the story.
Throughout my first data journalism class, we became familiar with the Washington Post's ARCOS database. For my final project, I acted as a local reporter in Clay County, Kentucky using the database for a story. I chose this county after filtering for counties that were majority white with a median household income of under $25,000. Of the 14 counties that fit that description, nine were in Kentucky. Further, four of these Kentucky counties, including Clay, received more than 100 pills per person.
I found that Clay County's population is 95% white, it received 137 pills per person and the median household income is $22,296. The Community Drug of Manchester pharmacy received more than 5.6 million pills, and it is now permanently closed.