• 

  Data visualization of ICE over the years

• 

  When in a zone

• 

  When user first is on

Inspiration

Trump during his second term made immigration one of his key policies. With this, ICE has gotten more power and funding. ICE has tripled its budget and now has daily quotas that need to be met for the number of people deported. With recent news in our local area and nationally about the way that ICE has brutalized citizens and undocumented people, we felt that we wanted to create a project that could help people stay safe.

Looking at 805UndocuFund as inspiration, we thought of a project that would implement citizen science and community support to inform impacted communities of ICE presence in an area. We thought of a general-use website that could be used by individuals to report cars that are highly suspicious of being ICE vehicles. With this report, it would populate a dynamic map that community members would be able to view, interact with, and verify.

How you built the project & challenges

We started this project recognizing that there were two main aspects to our application. The first aspect would be the front end, the actual website and information that individuals would use to report activity in their area. The other aspect would be the back end, making it possible for individuals to send photos to recognize license plates and providing a way for reports to be stored in a database and retrieved by others who access the website.

The front end had two main elements: the first would be visualizing the activity of ICE members in a person's community. The other element would be a visualization to inform the community about how ICE activity under the second Trump presidency is not normal and that the way ICE is operating requires individuals to inform their community of its activity. The map visualization was difficult, and the back end was not yet completed, so a minimum viable product of what it would look like on a person's screen needed to be made. We found that using street addresses was very difficult, as converting them into exact locations required APIs and other resources that were paid or resource-intensive. We settled on using the user's location and retrieving their latitude and longitude coordinates as a way for them to report where they saw ICE activity. This solution to a problem became a feature, as we realized it would also make it possible for users in the area to confirm whether ICE presence was still in the affected region. We then added a feature for users to click on the map to display their longitude and latitude coordinates in case they had already left the area and still wanted to report activity. The data visualization of how ICE has operated was difficult, as public datasets of these records are few and far between. Luckily, we found TRACReport, which is run by Syracuse University, that provided us with data. The problem we faced was that this data was not easily downloadable. We inspected the website, found the API where the data was being sent, and downloaded it as a JSON file. From there, we converted it into a CSV that allowed us to create a data visualization. We realized that showing data from such a long period of time proved to be cluttered and disorganized, so we found that adding filtering by year was especially useful for readability. For the backend, we needed a way to check license plates of photos that users would send in. That way, we could validate if the license plate belonged to ice using our existing database or add it to our database if confirmed to be a member of ICE. For this, we used a large dataset of license plates in the US including over 26,000 images of cars and license plates (provided by Fares Elmenshawii on kaggle). This allowed us to then create an object detection model under the YOLO family of object detection models to detect the location of the license plate in the image. After training the model with YOLO, we cropped the image to have just the image of the license plate and used another OCR model to get the text of the license plate. This, with good accuracy, allowed us to get the text of the license plate to then compare in our database.

What's Next

If we had more time and continued working on this project, we would try to host it ourselves, as we believe that storing data on Firebase, which is owned and operated by Google, may be hazardous for individuals whose location data is especially sensitive. We would also create two sides: one where an active community member could claim and manage a community to verify whether license plates are actually ICE, and another to make organizing and patrolling more efficient.

Built With

• css
• firebase
• html
• javascript
• yolo