Overview
What is police technology? How have they developed overtime? What surveillance technologies is the law enforcement in the US using? How can the law enforcement potentially harm vulnerable communities? What surveillance technology is implemented and used by state?
These are the following questions I will attempt to answer using Gephi by creating a network analysis graph.
After the death of George Floyd on May 25 of 2020, social unrest and protests proliferated all across the US and internationally. My social media feeds were permeating with history, information, updates, violence of police brutality in this country but also it being done globally. As community organizers and grass-root organizations began to share rallies and protests across social media, they also began to alert everyone to turn off their phones and dress discreetly because of drones and cell-site simulators. This was the first moment I consciously knew that I was one of many being targeted by police surveillance.
Process
Data
The Electronic Frontier Foundation partnered with University of Nevada’s Reynolds School of Journalism on Atlas of Surveillance: Documenting Police Tech in our Communities with Open Source Research project. The following features I was most interested were the following:
- City
- Country
- State
- Agency
- Type of LEA
- Summary
- Type of Jurisdiction
- Technology
- Vendor
Cleaning
The original dimensions of the dataset were 26 features by 7,303 observations. I removed all observations where Agency, Vendor ,Technology, State were undefined. I uploaded my dataset into OpenRefine to update the discrepancies of naming conventions to make sure they all coincided.
Preparing
My initial goal was to use the features laid out above, but I experienced difficulty when setting my edge table in Gephi. Somehow, it duplicated my target column onto the nodes Id column, creating an insurmountable number of null values when beginning to graph. Because of this, I used the following video Convert Excel CSV into Network in order to make my graph more understandable to me
Edge Table
Target being Technology
Source being the State
Type being Undirected
Weight
Nodes Table
Id & Label were State and Technology
Frequency
Type: Source or Target
Visualization
As the following visualization points out, the surveillance technologies registered in the dataset are as follows:
- Body worn cameras
- Automated license plate readers
- Camera Registry
- Cell-site simulator
- Ring/Neighborhood partnership
- Drones
- Face recognition
- Video analytics
- Real-time crime center
- Gunshot detection
- Predictive policing
The network analysis shows the relationship between the above surveillance technologies with states in the U.S. I found it interesting takeaways were Puerto Rico and the US Virgin Islands are only utilizing gunshot detection surveillance. Florida has the highest use of facial recognition technology, Tennessee has a strong use of predictive policing technologies.
