Application of Data Analysis in Trend Prediction of Different Crime Types in London

Abstract

This paper aims to analyze the data of different crime types in the London area for the past three years. The data is from data.police.uk provided by City of London Police. The R language and ARIMA model were applied to forecast different crime types in London. Seven sets of stable non-white sequences were obtained by preprocessing the original data, stationary test, and non-stationary sequence differential processing. The time series analysis method was used to realize the ARIMA model modelling. After comparison, the prediction was performed with the optimal ARIMA model. In terms of anti-social behaviour (i1), public order (i9), robbery (i10), shoplifting (i11), theft from the person (i12), the predicted results show a small fluctuation. And it shows a large fluctuation trend in bicycle theft (i2), but the overall number is lower than the previous months. Violence and sexual offences (i14) show a downtrend.