Update: We won first place! Thanks to the competition organizers and everyone involved

2023 IMI BIGDataAIHUB Case Competition

Alexandre Granzer-Guay, Jesse Ward-Bond, Muhammad Maaz

Repository here: https://github.com/jwardbond/2023-imi-bigdata

This repository contains our submission for the 2023-2024 IMI BIGDataAIHUB Case Competition. In this competition we were given synthetic bank transaction data (emts, wire transfers, cash transactions, and customer data) and given three tasks:

1. Build a simple classifier to flag high risk customers (money laundering)
2. Identify wildlife trafficking networks in the data using graph-based techniques
3. Scrape trafficker names from online sources match them with our data.

We've completed all of these tasks, and our approaches to each of the - and the webapp GUI we developed to explore our results - are contained in the respective folders within this repository.

Setup

1. If you don't have jupyter installed already, follow this guide
2. pip install -r reqs_webapp.txt if you just want to run the GUI OR pip install -r reqs_full.txt to include dependencies for running the notebooks.

Running the webapp

1. Install the requirements above.
2. Navigate to the webapp folder: cd .\webapp\
3. Launch the flask server using python app.py.
   • It will take around 20 seconds for the server to construct the graph.
4. Once the flask server is running (you should see Running on http://127.0.0.1:5000 in the console), open the index.html file in any web browser.

For additional information, consult the README.md within the .\webapp\ folder.

Running the notebooks

• All notebooks should be run top-to-bottom, in the order they are stored presented in their respective folders
• We have cached all results, but if for some reason you want to re-run everything, it would be best to run the tasks in the following order: task_3 > task_2 > task_1
• The notebooks are well-documented, so README's are not provided within .\task_1\ etc.