Predicting Graffiti using City Layouts

The Task

This project will analyse what properties of a city correlates with graffiti generation in Vancouver, CA. Cleaning costs for graffiti can be a significant expense in a big city, and reducing it is in the interest of the city council of Vancouver as well as private companies. Therefore, we will be aiding governments to allocate efficient and effective budgets with regard to managing graffiti around their cities.

Necessity

In July 2021, Vancouver city spent half a million dollars on graffiti removal. During the Covid-19 pandemic, there was a spike of nuisance graffiti, resulting in an increase of 87% in calls to 311. The presence of graffiti in an area over a long period of time encourages the appearance of other graffiti. Moreover, graffiti has sometimes led to racist attacks. Lastly, this problem goes beyond just Vancouver. It is a global problem. On the other side of the globe, Australia spends 1.5 Billion Dollars annually on graffiti removal.

The Goal

The aim of this project is to help predict graffiti based on features of the city, making it easier to focus cleaning efforts. We will focus on individual buildings, and using the provided data on how much graffiti each building has, we will be able to identify features which led to the building being clean or vandalized. This project will help the city council of Vancouver improve the livelihoods of their citizens by having a cleaner environment. Moreover, the data will be used to prevent further graffiti by suggesting city layout changes or increased enforcement in select areas where this is needed. It will help the city council of Vancouver to act earlier and more precisely if there is a potential for other graffiti to appear in some areas. The project will also be able to identify which common features of cities prevent or encourage graffiti to be put there in the first place.

Data Sources

Data from the Open Data Portal of the City of Vancouver will be used in this project. We use a total of 8 datasets from this repository.

Target Audience

This project is primarily aimed at the City Council of Vancouver, since we chose Vancouver as the location for our study. It is publicly available for other officials from around the globe to learn from. We hope this research will prompt a new way of dealing with graffiti before it occurs. For example, city planners may use the information to design graffiti-proof cities and mitigate or solve current graffiti related problems. More technical details about the methodology can be found in our report on github, with the hopes of inspiring other researchers to explore this area further.