Inspiration
Despite us being Americans crime in Canada still resonated with us. The crime data of Montreal reminded us of the tragedy of recent mass shootings in the US, and with the prevalence of school shootings on the rise, as students we wanted to discover insights within the Montreal crime data in order to prevent further tragedies and help save lives.
What it does
Using data visualization graphs we were able to condense unreadable raw data into a easily digestible format. Our graphs reveal things like crime hotspots in Montreal as well as the type of crime that occurred in that location. We were also able to create a regression algorithm that allows us to input a Montreal address and get out a ranking of each type of crime to from most likely to least likely happen at that given location.
How we built it
We decided to use python along with the Matplotlib, Seaborn, Pandas, and Geopandas libraries in order to support our data visualization. We used pandas to extract raw data, Matplotlib plot the graphs and Scikit to create a regression model to predict which types of crimes are most likely to happen in each district. We also use Nominatim's API service to convert the given address into lat/long coordinates.
Challenges we ran into
As a group that is entirely new to data science projects we ran into many challenges. Due to it being our first datathon we weren't used to working for so long and often struggled to keep our energy up. However with good time management and plenty of breaks we were able to push through to completion.
The largest problems we faced were related to data integration, and the time constraints of the project. We found it difficult to relate the two given data files in the Montreal crime prompt, but by breaking down each file piece by piece we were able to combine the two data sets and create one visualization graph that combined the information from both sets. The time constraints were also an issue since we started off relatively slowly. We eventually had to cut some of our initial goals down to realistically finish by the deadline.
Accomplishments that we're proud of
Being new to data projects we found accomplishments every step of the way. From basic things like reading the data files and separating the data into meaningful groups to more complex things like combining data sets to find relationships between them and creating our prediction algorithm. All in all, despite all of the struggle, we're very proud of our finished project and are very happy to showcase what we have!
Log in or sign up for Devpost to join the conversation.