S.E.B — Software Engineering with Barrett
Labs 32 assigned my team the ‘Human Rights First — Blue Witness’ project. This site looks through twitter and reddit for instances of police use of force on peaceful people. Having multiple sources of reported incidents, this application removes duplicated incident reports and consolidates reporting on Police use of force. With that data, a map is populated with incidents and locations, a timeline of recent incidents is displayed further down on the home page, graphs show different incident trends, and an admin user can approve new incidents and activity for the database.
Inheriting a project and the challenges that come with it
Having inherited a project from a previous team, there were many technologies implemented that are not necessarily best practice. This caused compatibility issues, as well as a learning curve for the team. For some team members, they have never seen these technologies implemented in a project. However, through collaboration, team members who had worked with these programs were able to guide the team members in learning and implementing these technologies.
Once these issues were addressed, the team was able to assign tasks to team members and focus on mission accomplishment. The changes to the code base were primarily styling and responsiveness. Once deployed, not all changes were present. This arose due to using .SCSS. Some work was done on the back end working with the data science department. There were conflicts with the database structures, preventing new incidents to be pushed to the database.
The journey of 1000 steps
After the team worked relentlessly on the product, we have brought a lot of changes to the deployed site. We have added many UI changes including footer, link to parent site, and admin access. To consolidate information, we moved all graph data to a single page and implemented a message if no data is available for searched locations or dates. On the back end there is now capability to delete incidents that were mistakenly pushed to the database. The data science team resolved many issues, some of which were issues of assuming that tables were already initialized, or they ad data, and starting off some aspects as notebooks. Data Science implemented deduplication methods to prevent multiple tweets from being published. Finally they made it possible for twitter data to be merged into approved incidents.
The future of this project is bright. There are still some stylistic updates that need to be implemented, however a team will be able to achieve these with standard effort. As of now admin users still have manual editing of incidents. This can be automated with machine learning and more data science engineering. We have got the site functioning from our data base, and all required features have been implemented. This project is important not only for us as developers, but will provide real world services. I am excited to see this project go public!
-Christopher Barrett