Stoplifting

100 Shoppers and their paths taken with the security guard in blue.

Inspiration

Living in the UK, we realised that inflation led to not only an increase in the cost of living but incidences of shoplifting we observed day to day. Delving further into the subject, we realised how this issue was a vicious cycle and the impact it had on our day-to-day lives.

What it does

We have created an Agent-Based Model which is able to simulate interactions within grocery stores with security guards and shoplifters. This allows us to predict the performance of security systems used in grocery stores, and identify the best methods to reduce shoplifting. The data produced can then also be used for training an ML model, which can provide powerful information to grocery stores that aim to reduce shoplifting as much as possible.

How we built it

We leveraged the powerful IBM Z platform, collaborating to create our proof of concept Agent-Based model which is robust and can be re-tuned easily for future improvements.

Challenges we ran into

There were difficulties in modelling the behaviour when multiple Agents interacted on the same grid tile at the same timestep. There were also some inefficiencies in the code that were unable to be ironed out in time due to the Datathon being only 24 hours. However, with IBM Z Cloud Computing, these inefficiencies did not prove to be problematic at this stage and the simulation was a success.

Accomplishments that we're proud of

Learning how to use IBM Z Cloud Computing Building the simulator, and attempting to build an ML model with the results Teamwork Kindness Friendship Cohesion

What we learned

How powerful the combination of our dreams and skills, combined with IBM Z Computing are.

What's next for Stoplifting

Our dream would be to partner up with large grocery stores in the UK. This would allow us to access data which we are having much trouble accessing now, and truly limit-test the abilities of our model and harness the full capabilities of IBM Z Cloud Computing.