Resubmission of Hackathon Project Canadian PR

[pibeebeks]

Project name

Resubmission of Hackathon Project Canadian PR

Description

This project designs an end to end solution for Canadian Permanent Resident Applicants with Microsoft Fabric. The project facilitates the analysis of quantitative and qualitative data from the Immigration, Refugees and Citizenship Canada (IRCC) website, predicts scores for Canadian Permanent Resident draw scores, leverages on RAG architecture via a chatbot interface, and visualizes the data using PowerBI.

The data used for the project originated from the Immigration, Refugees and Citizenship Canada (IRCC)
https://www.canada.ca/en/immigration-refugees-citizenship

Overview

The model performs the following tasks:

1. Data Collection and Storage:

   • Retrieves quantitative and qualitative data from the IRCC website.
   • Utilizes OpenAI GPT-4 to analyze the qualitative data.
   • uploads these both quantitative and analysed qualitative data to a Fabric Lakehouse
   • Automates the process through a data pipeline and updates the lakehouse weekly with new data.

2. Prediction Model:

   • Feeds quantitative and qualitative evaluation data from the Fabric Lakehouse into a Synapse prediction model.
   • Predicts scores for the PR draw across different classes.
   • Stores these predictions into the Fabric Lakehouse.

3. Chatbot Interface:

   • Develops a retrieval augmented generative model with OpenAI GPT-4.
   • Retrieves quantitative and qualitative data from the Fabric Lakehouse.
   • Answers various queries through a chatbot interface.

4. Data Visualization:

   • Utilizes PowerBI to visualize qualitative data, quantitative data from IRCC, and predictions from the model.

Project Repository URL

https://github.com/Pelumioluwa/Microsoft-Fabric-Hackathon/tree/main

Project video

https://www.youtube.com/watch?v=870XSyqM0gA

Team members

Pelumioluwa - https://github.com/Pelumioluwa