Improviz
Exploring the crimes in Chicago from 2001 to 2019
Chicago is the third most populated city in the United States with almost 3 millions resident in 2019. The city is located in the State of Illinois, and is classed as a "Alpha world city" the Globalization and World Cities Research Network
(GaWC).
The Chicago Police Departement started to record the crimes since the year 2001. An astonishing observation is that the dataset containing all the crimes from 2001 to 2019 has more than 7 millions entries!
In this project we present the datas and try to show the relationship/link between the number of crimes and some socio-economics factors, and the locations of police stations.
Here is the first insight: the mean number of the total number of crimes of all types for each days of the year.
We see that the average number of crimes are generally constant over the observed years. But there are some exceptions!
The most obvious one is the big peaks of crimes the first of January. The parties the night before the new year's eve must have something to do with it.
Something surely has caught your attention as well: there are peaks of crimes for every first of the month! We have no clear explanation about them, only some suppositions.
The first one is the naive one: the first of each month has an impact on crime, maybe due to people receiving their Government welfare and Social Security checks or thinking the beginning of a new month is the occasion to do crazy stuff. But surely, even if the
dataset is provided by a big city, we believe that those peaks comes from
errors in the entries and that if there is no precise date for a crime, the value is defaulted to the first of the month.
Finally, we can see that the day with the lowest crimes rate is christmas, which could be explained by the fact that the majority of the population of the United States is christian.
On this streamchart each stream is a category of crimes.
This second display presents the proportion of crimes by their type. This shows that the majority of crimes are part of the less severe type, with Theft, Narcotics, Criminal Damage and Battery being the most represented. And it is reassuring to
see that crimes such as Homicide, Criminal Sexual Assaults and Kidnapping are really thin lines, making them a minority of crimes which have been registered by the Chicago police. More a bit more common are the Burglary, Robbery and Motor
vehicle theft which we could have expected to be even higher when only seeing the total number of crimes registered.
This gives a nice overall idea of what can be expected when using the filter mode in the map section
The interesting part with those reported crimes is that the data is also spacial, so now let's see on a map what those crimes look like! Hint: don't be shy to press the buttons on the right.
We can observe that crime density doesn't seem to be affected by police stations proximity. But is more linked to the geographical location in Chicago, for instance we see that the density is high in central Chicago. More specifically in Near North Side, which is the northernmost part of central Chicago and the Loop which is the main section of Downtown Chicago.
Another interesting point is that the crime density tends to be lower the more we get close to the present day. This could be due to a first adaptation by the police force, which could dispatch more efficiently it's manpower by using the data it collects.
This brings us to the next point, in which some socio-economic statistics can be observed on the Community Areas which interest us.
At this point we can start our assumptions based on socio-economical statistics. Using the previous section we chose to focus on two small areas. Near central Chicago, with the Community Areas the Loop and Near North Side, which had a high crime
density. And the West Side of Chicago, with New City, Bridgeport and McKinley Park, which have in comparison a very low crime density.
The first thing that we observe is that in central Chicago the unemployed percent and adults without high school diploma is lower. By observing the crime distribution, it shows that Theft is the main problem, while in the West Side, the
distribution is more evenly accross Criminal Damage, Theft and Battery. We suppose this is due to having less "opportunities" for Theft in the West Side since the households below poverity are lower in that area.
This gives us a first overview on relationship between socio-economic statistics and crime density and distribution. The aggregation of data and visualisation helps a lot when trying to understand relationships between them, and being able to get
parts of information from every sections is really helpful.
This is the end of this data journey... So what's next ? We have a couple ideas of interesting features that could be done with those datas.
Machine learning algorithms can be used to predict the crimes on the city in order to help the police to place the patrols accordingly.
Even more, a path recommander that propose the safest routes to go from a point to another could be done.
