These R Shiny scripts and re-formatted data files are in support the amazing project, Who's Making News: https://www.whoismakingnews.com/
The initial intent of these scripts was to (1) teach myself Shiny, and (2) make some interesting analyses/visualizations for looking at data and support the cause. On the Shiny page, I bring in other data sources and work through the following analysis methods:
- Linear regression to predict today's crime rate.
- Time Series decomposition to predict today's crime rate.
- Exponential Smoothing to predict today's crime rate.
- Group by category and sub-category for data exploration.
- Breakdown of criminals by gendered name into "male" names, "female" names, and uncommon names (to play with word clouds).
- Compare crime rates by state in terms of parameters such as demographics, employment, policy and politics by linear regression.
- Compare relative crime rates by state to other crime statistics using Principal Component and K-means Clustering.
Other standard analyses of these data could be found elsewhere (rates per capita, etc.), but what struck me as interesting after a quick look through the data was the preponderance of male names associated with sex crimes against children. I wanted to be able to emphasize this information and play with how to best represent this.