Project 2 – Week 4

This week, I explored the relationship between three types of political violence events normalized per 1 million people:

• Protests_per_1M
• Riots_per_1M
• Violence_per_1M

Using .corr() in Python, I calculated the pairwise correlation coefficients. All values were positive, indicating that states with high values in one type of event tend to have high values in others as well.

I created three scatter plots to see how the number of protests, riots, and violence events per million people are related across states. All three plots show a positive relationship — meaning when one goes up, the others also tend to go up.

Variable Compared	Pearson r	Strength	Comments
Protests vs Riots	0.78	Strong	Strong upward trend, many points clustered near line
Protests vs Violence	0.89	Very strong	Tight clustering, highest correlation among the three
Riots vs Violence	0.71	Moderate–Strong	Still positive, but more scattered

The first plot shows Protests_per_1M vs. Riots_per_1M, pearson r = 0.78. This plot shows that states with more protests usually also have more riots. The relationship is quite strong, meaning that when protests go up, riots often go up too.

 

The second plot compares Protests_per_1M vs. Violence_per_1M. There is a very strong link between protest activity and violence against civilians. States with higher protest rates tend to see more violence happen to people, and this pattern is very clear in the data.

 

The third plot shows Riots_per_1M vs. Violence_per_1M. This plot shows a noticeable connection between riots and violence. While not as strong as the other two, the trend still suggests that more riots are often followed by more violence affecting civilians.