Overview

The Project

This project seeks to find any significant correlation between the Great Migration and the emergence of African American film companies in the United States. I wanted to find out if the Great Migration, in particular, caused a rise in the establishment of film companies that produced race film in the northern half of the United States, which was profoundly impacted by the Great Migration.

In order to find this, I used an Early African American Film data set as well as US Census data for population increase between 1890-1900, 1900-1910,  1910-1920, and 1920-1930. I used these dates to stay within the parameter of the course guidelines, and because the African American film companies seen in the data set I use were founded within these dates. In this project, I analyze the population increase data with the founding dates and locations of these race film companies to come to conclusions about their possible correlation.

The Data

Early African American Film

The data set which inspired this project comes from the collaboration of various graduate and undergraduate students of UCLA’s Digital Humanities program. The data compiles a list of production companies, actors, and films included in the race film’s silent era. For this reason, the data set chooses to only include films prior to 1930. For my purposes, I use only the production company aspect of the set.1 Within the set, I chose to further narrow my focus and only include film companies that: included whether they were African American or white-owned, included the founding date, and included the founding location. Due to my narrow focus, I took the data set from 175 rows of data to around 60 rows of data.

The challenges with this data set are that the sample is very small due to a lot of information missing, either from the data set creation or from history as a whole. Many companies in the data set either did not have a founding date provided, the owner of the company was unknown, or the foundation location was unknown. However, even with these limitations, I still believe I have a good snapshot of the historical trends this data set wants to represent.

US Census Data

I used US Census data from the 1900, 1910, 1920, and 1930 US Censuses. These dates were chosen, because I believe they best match up with the Early African American Film data set previously discussed. In particular, I chose to look at the Population Increase tables within each census. This is because I wanted to see the population as a reflection of the Great Migration which was occurring during this time.

The challenges I faced using the census data made this project very time consuming, and often frustrating. The process used from finding the data to converting the PDF files to CSV files, then cleaning the data made this part particularly difficult and time consuming. Not only that, but I found that the table names and language used are not uniform across census years, making it a little difficult to look for and find exactly what was needed for this project.

Methods

Choosing Data

I chose to use the Early African American Film data set, because this data set combines my two areas of study: history and film. Also, I find the race film genre particularly interesting as this was not something I had learned before this semester (both in HST489 and FLM300), and I think it is a really important area of film and African American history to explore.

Cleaning Data

To clean the data, I used Microsoft Excel to manually extract the information I needed and to clean anything that was converted from PDF to Excel. To save and convert data, particularly census data from PDF to Excel, I used an app called iZip to store and extract zip files and one called PDF to Excel on my iPad Pro. Some complications from this is that the PDF reader read some of the superscript numbers as digits for population, and some of the pages had numbers that were difficult to decipher for the PDF reader. For the last two census data sets, I converted the information I needed manually.

Visualizations

I chose to rely on mapping for my visualizations. For both the census data and film company data, I used time maps. When interactive, the census map show the change over time for population increases, and the film companies map displays the locations and demographics based on the year each company was founded. For my visualizations I used the kepler.gl program with CSV files created in Microsoft Excel.

I struggled with the visualizations. For the film companies, I was originally interested in a time map to see change over time with the emergence and closing of the various film companies. I was unable to fulfill my complete goal due to the data set missing a lot of closing dates for the film companies, and had I included only those with closing dates, the sample size would have been much too small. For this reason, this time map feels a bit incomplete, but I still feel as though it provides a good glimpse into the foundation of race film companies. I was interested in creating geoJSON file for the census data, but could not figure out how to make this work due to time restrictions. Additionally, I couldn’t figure out how to place my interactive maps into this website. For that reason, the visualizations are screenshots of the interactive map.