After starting with the ACLED-India dataset, I decided to shift to the ACLED-USA dataset based on class discussions and the professor’s suggestion. The U.S. dataset feels more focused, with cleaner data and a strong emphasis on protest-related events, which makes it easier to work with and better suited for the kind of analysis we’re learning this semester.

This week, I spent time getting familiar with the structure of the dataset—looking at different event types, who was involved, where events happened, and whether there were any fatalities. A lot of the events involve civilians, law enforcement, and militia groups, which opens up several interesting directions to explore.

I’ve started outlining a few key questions I might focus on:

• What kinds of protests tend to turn violent?

• Are certain states or areas more likely to experience unrest?

• Do protest patterns shift around major political events like elections or court decisions?

• Are some serious events underreported based on where they happen?

Next, I’ll clean the data and start experimenting with clustering and pattern analysis to help answer these questions.