Tuesday, 3 March 2015

Analysing Crime Data

The Information School’s MSc Data Science programme aims to give students the chance to enhance their analysis skills by working on real data sets.  As part of a recent assignment, students have analysed data sets on crimes which have been reported to the UK police.

One assignment, conducted by Rebecca Thorpe, investigated burglaries in the South Yorkshire region and specifically examined where burglaries occur in the Sheffield area, also looking for associations with other variables.  Using boxplots, time series, line charts and Holt-Winters forecasting, this assignment found that the mean number of monthly burglaries in South Yorkshire was higher in 2012 than in 2011 or 2013.  Using Holt-Winters forecasting, the assignment predicted that burglaries would decrease in South Yorkshire over the next three years.  Focusing upon burglaries in Sheffield, it was found that instances were highly concentrated in the eastern areas of the city and there was some correlation between unemployment and the number of burglaries.

The following diagrams were included in the analysis which was conducted in this assignment.

Box-whisker plot showing total number of burglaries in South Yorkshire per month from 2011 to 2013:



 Burglaries per capita in Sheffield from 2011 to 2013:



A second assignment carried out by Joseph Ellard analysed instances of anti-social behaviour in Cambridgeshire between 2010 and 2014 and also analysed bicycle theft levels.  Analysis was carried out using R, Holt-Winters forecasting, linear regression and ggmap for plotting.  Key findings from this assignment were that crime levels peaked during the summer and that they had decreased between 2010 and 2014.  Analysis also found that warmer weather typically resulted in an increase in crime, while increased rainfall caused fewer instances of crime.  Bicycle thefts were also found to peak during summer months although they were highest during September and October.  Most bicycle thefts were found to occur around Cambridge and Peterborough.

The following diagrams were included in the analysis which was conducted in this assignment.

Anti-social behaviour correlations:


Bicycle thefts in Cambridgeshire:


These excerpts from student assignments highlight the skills and techniques that are taught on the MSc Data Science course which prepare our students for a career in the data science industry.   For more details about the course please visit our website.




No comments: