I used data from data.police.uk, a site that provides open data about crime in England, Wales, and Northern Ireland. I had initially wanted to use data from the Metropolitan Police, Greater London’s police force, but the amount of data was immense and CartoDB was unable to connect to a dataset of such multitude. In lieu of the Met’s data, I used data from the City of London police. The City of London, also known as the “Square Mile” because it is 1.12 sq miles/2.90 km, is a a hub of financial institutions and home to a relatively modest population of ~520,462 people. I downloaded a bundle of Excel spreadsheets, which contained crime data from December 2010 to February 2015. Since each month had its own Excel sheet, I aggregated all the data into one sheet, then connected the dataset to CartoDB.

The first viz I created in CartoDB is “Crime in the City of London (December 2010-February 2015).”

[Click image to access interactive version].

This category map visualizes drugs, shoplifting, burglary, other theft, vehicle crime, anti-social behavior, violent crime, violence and sexual offenses, criminal damage and arson, and miscellaneous crimes (“other crime”)  as documented by the City of London police from December 2010 to February 2015 (which is the most current data available). There is a key on the bottom right-hand corner which a user can use to differentiate types of crime (by color). If the types of crimes are not closely related I tried to make sure the colors were distinct enough in order to more clearly distinguish them from each other. If they are related (like shoplifting, burglary, other theft and violent crimes and violence and sexual offenses) I placed them in the same color family. I also included a hover feature for convenience, so a user can immediately know where a crime was committed instead of zooming in to find out–a user can hover over a point and be shown where a crime was committed “on or near” a particular street, landmark, or highway.

This viz could be improved if I were able to create a more suitable UI–a filter could be included in order for a user to view crime in a particular month/year (though,  I created an animated heat map [see viz 2] for this purpose). I also think displaying the numerical amount of crimes for each point could enhance the visualization.

Lastly, overall, one can immediately glean from this viz that there have mostly been “other” thefts, anti-social behavior, and “other” crimes (I’ve provided a list of what types of crimes are included in “other” at the end of this post).

The second viz, entitled Crime in the City of London Over Time, is an animated heatmap version of viz 1.

This viz shows the amount of crimes over time (December 2010 to February 2015). While it may look like an amorphous blob at first glance, there is a method to the madness! Red indicates a greater intensity  i.e., a greater amount of crimes. Zooming in allows a user to better track the amount of crime over time for a specified location. The time slider, automatic on page load, allows a user to pause the map on a particular month/year in order to better view the amount of crime for that month/year or to manually slide across the months/years. Since the time slider plays automatically, I have slowed down the animation (I’ve set the duration [secs] at 40) just enough so that a user is able to compare the increase/decrease of crime between consecutive months.

Again, this viz could be improved for better analysis purposes with a filter by types of crime. But, mostly, this map shows there has been some consistency of crime in relation to particular locations. It also provides some other insights such as, for example, 2011’s progression–October 2011 is suddenly far more intensely red than previous months, indicating an increase in crime.

The last visualization concerns violent crime.

 

I thought it would be interesting to isolate one crime type, so I’ve filtered type of crime by violent crime and violence and sexual offenses and created the above viz. With cluster visualizations, is it possible to see exactly how many crimes were committed in a specific location. While it is not obvious at first glance which location corresponds to which circle, a user can zoom in and read the street names. As a user zooms in, some circles break and become more granular so a user can pinpoint a more exact amount at a more exact location. I do think this viz could be improved if there were a time slider to track the growth of violent crime in particular areas.

As a final note, I think CartoDB’s search feature (in the upper right-hand corner of each viz) is quite useful; if a user wants to check crime that has occurred near his/her residence, he/she can type in the street they live on and magnify it.

---

Police.uk’s FAQ provides a list of definitions for each type of crime in the visualizations above, which clarifies some of the more vague labels for the types of crime:

• Anti-social behavior: Includes personal, environmental and nuisance anti-social behavior.
• Burglary: Includes offences where a person enters a house or other building with the intention of stealing.
• Criminal damage and arson: Includes damage to buildings and vehicles and deliberate damage by fire.
• Drugs: Includes offences related to possession, supply and production.
• Other theft: Includes theft by an employee, blackmail and making off without payment.
• Robbery: Includes offences where a person uses force or threat of force to steal.
• Shoplifting: Includes theft from shops or stalls.
• Vehicle crime: Includes theft from or of a vehicle or interference with a vehicle.
• Violence and sexual offences: Includes offences against the person such as common assaults, Grievous Bodily Harm and sexual offences
• Other crime: Includes forgery, perjury and other miscellaneous crime.

---

Sources:

1. http://www.data.police.uk

2. http://www.ukcrimestats.com/Police_Force/City_of_London_Police

3. https://www.police.uk/about-this-site/faqs/#what-do-the-crime-categories-mean

 

Sarah Hatoum