Forecasting London's Future Crime with ARIMA

Data mining is like digging for treasure in a huge pile of information. The London police noticed a sudden rise in crimes across different districts in 2017. They had no clue how to predict future crimes. To solve this, a team used a time series model called ARIMA to forecast crime rates in London. They fed ARIMA with five years of crime data and asked it to predict the next two years. ARIMA did a better job than other methods like exponential smoothing. The data was gathered from the London police website and other sources.

The project had four main steps. First, they extracted crime data from web sources between 2012 and 2016. Next, they cleaned and structured the data. Then, they used IBM SPSS to visualize the model. They also calculated moving averages, differences, and auto-regressions to make predictions. The model was quite accurate, giving correct predictions 80% of the time. This helps London police make informed decisions to fight crime.

Actions