Australian-Rental-Insight-Copilot

A Streamlit app for exploring Australian rental market datasets with automated statistical analysis, rental-focused visualisations, multivariable plotting, and anomaly detection.

Overview

This project is designed around real Australian housing data rather than a generic CSV dashboard. It focuses on practical rental-market questions such as:

• Which suburbs are more affordable for renters?
• How do bedrooms, property type, and location affect weekly rent?
• Which listings or postcode groups look unusually expensive relative to the rest of the market?

Real dataset strategy

The app is structured to work with two stronger public datasets:

Primary dataset

• Australian Rental Market Data 2026

Secondary dataset

• NSW Rental Bond Lodgement data

This gives the project:

• a national cleaned rental listing dataset for broad EDA
• an official NSW government dataset for validation and trend-style analysis

Features

• CSV upload with automatic column-type detection
• rental-focused field inference for rent, suburb, bedrooms, bathrooms, property type, and postcode
• summary statistics and missing-value analysis
• filtering by suburb, property type, bedrooms, and rent range
• histograms, box plots, suburb comparisons, and multivariable scatter plots
• correlation matrix for numeric features
• anomaly detection using IsolationForest
• plain-English insight summary generated from dataset statistics

Tech stack

• Python
• Streamlit
• pandas
• Plotly
• scikit-learn

Repository structure

.
├── app.py
├── requirements.txt
├── data/
│   ├── external/
│   └── sample_rental_listings.csv
└── src/
    ├── __init__.py
    ├── analysis.py
    ├── dataset_adapters.py
    └── schema.py

How to run

git clone https://github.com/aadyasingh55/Australian-Rental-Insight-Copilot.git
cd Australian-Rental-Insight-Copilot
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
streamlit run app.py

External dataset setup

Place one or both real datasets in data/external/ using these filenames:

data/external/australian_rental_market_2026.csv
data/external/nsw_rental_bond_lodgement.csv

The app will automatically:

• prefer the Australian Rental Market dataset if present
• otherwise fall back to the NSW bond lodgement dataset
• otherwise use the bundled sample CSV as a backup

Dataset expectations

The app works best with rental-style CSV files containing columns similar to:

• weekly_rent
• suburb
• bedrooms
• bathrooms
• property_type
• postcode

The bundled sample CSV exists only as a fallback so the app can still run without external data.

Current scope

This repository is an MVP analytics tool rather than a full production rental intelligence platform.

• It works best on relatively clean tabular rental listing data
• Schema inference is heuristic, so some real-world exports may need minor cleanup
• The anomaly detection and insight summaries are exploratory signals, not final market judgments