Narratives from Numbers: How data and analysis can help you to tell stories
Data without context don’t tell us enough: they might tell us how something is, but they won’t tell us why that is or what can be done to change things. Good analysis can help you to identify the factors that produced a situation and can tell you what needs changing and how.
Last year Get the Data was commissioned to contribute to a multi-agency report for the Mayor of London’s Violence Reduction Unit (LVRU), Neighbourhoods affected by violence. Using data from the Metropolitan Police (the Met) the report looked for trends in violence in London – where violence was most likely to happen, what factors were driving violent crime – and made recommendations for city-wide approaches to tackling violence and how to collect and analyse data more usefully.
By themselves the Met’s data are simply a record of the numbers of violent crimes in London over a three year period. But with the proper analysis these numbers can be made to tell a story: a story about what drives violence, who suffers from it, and how these sorts of incidents can be reduced in number.
I worked with colleagues from Get the Data, taking the raw data from the Met and analysing it using a variety of methods, to show where the majority violent crimes are happening and to identify the social conditions influencing those trends.
A high proportion of violent crime clusters in small areas and to get the most detailed information the data need to be analysed at a low level of geographical detail. These areas (of about 1,500 residents) are defined by the Office for National Statistics as Lower Super Output Areas (LSOA).
We started by choosing a metric – average yearly rate per 1,000 residents – in order to standardise the results, and then calculated the results for each LSOA. Once we had done this we could then correlate these clusters with factors such as health, education, the night time economy of the area, and geographical location within London.
Our analysis used multiple analytical approaches to prevent duplication of results, to remove factors that may have been caused themselves by the rates of violence, and to identify outlying data that may have skewed the overall results.
One analytical challenge was how to deal with nighttime economy areas, where the violent crime is brought to the area rather than emanating from there (the latter was the LVRU’s interest). For example, Westminster, a wealthy borough in the centre of London, had some of the highest rates of violence. Our analysis quickly established that this was due to an influx of visitors to the area due to the nighttime economy; the violence was being done by people from outside of the borough. Our analysis of multiple factors identified the likely nighttime economy areas and they could be placed in their proper context so that a more accurate story could be told.
The results of the analysis clearly show that factors associated with social exclusion and economic deprivation correlate strongly with a higher risk of violent crime. Identifying which social factors function as long-term drivers of violent crime opens up a variety of preventative policy options and targeted interventions to reduce violence alongside the more obvious reactive strategies.
By using multiple approaches to analysing the raw data a story began to emerge about where violence is most likely to occur and what factors are associated with a higher risk of violent crime. By knowing this organisations working to reduce violence in London and to mitigate its effects now have stronger evidence for which policies are likely to achieve this, and where to target their scarce resources for the best outcomes.
At Get the Data we’re passionate about helping you to get the most from your data. If you or your organisation want help to tell your story, to improve outcomes and target organisational resources more efficiently, then please contact either Alan or Jack or visit getthedata.net/contact.
