A cloud-focused machine learning demo with a FastAPI backend and Streamlit frontend for UK housing predictions, plus an electricity endpoint added to demonstrate the full app pipeline.
This was a group project by a team of 3 and is best presented as a full-stack ML deployment demo. The repository combines a FastAPI backend, a Streamlit frontend, model prediction endpoints, and an AWS-oriented workflow for notebook execution and data handling.
It demonstrates how machine learning can be wrapped into a usable product experience rather than staying as a notebook-only experiment.
The strongest and most defensible part of the project is the housing price prediction flow. The repository states that this model was trained and tuned with PyCaret and then selected as the deployed model.
The electricity route is useful for showing interface and integration work, but it should be framed as part of the app pipeline demo rather than as a standalone forecasting system.
2025
Cloud-focused ML deployment demo
UK housing prediction with deployable app flow
Group project by a team of 3. We worked across the model, API, frontend, cloud workflow, and integration together.
VS Code, GitHub, FastAPI docs, Streamlit
Prediction endpoints are exposed through FastAPI, with a docs interface that supports testing and integration.
Provides the user-facing app with separate flows for housing price and electricity demand prediction.
The project references S3 and SageMaker, while the housing model was trained and tuned with PyCaret.
Trained and tuned the housing model with PyCaret as the main predictive component.
Wrapped prediction logic in FastAPI endpoints for testing, integration, and deployment-style use.
Built a Streamlit interface to make both prediction routes accessible through a simple app experience.
Connected the project to an AWS-oriented workflow using S3 and SageMaker concepts in the pipeline.
