Cracking Crime with Time Series

Ever thought about predicting crimes before they happen? Well, that's exactly what the London police are trying to do. Starting in 2017, crimes in London's different boroughs started to rise rapidly. But how do you stop something that's already happened? That's where data mining comes in. By digging into large datasets, we can find patterns that might help us predict future crimes. One handy tool for this is the ARIMA model. It's like a clever detective that uses time series data to solve the crime puzzle. We fed the model five years' worth of crime data from London and asked it to predict the next two years. Guess what? It did a better job than the other detectives using exponential smoothing. The data came from real crime reports on the London police website and other resources. We broke down our approach into four parts: data extraction, data processing, visualizing the model in IBM SPSS, and finally, making predictions. Data extraction pulled in crime data from web sources between 2012 and 2016. Data processing cleaned and organized this data, giving it specific attributes. The model then crunched the numbers, calculating moving averages, differences, and auto-regression. The result? An accurate model that got things right 80% of the time. This isn't just about numbers; it's about helping the London police make better decisions to fight crime.