Apartment Matcher 🏠🔍

A web application that allows users to filter apartments for buying or renting and utilizes a Machine Learning model to find the best apartments based on user preferences.

Table of Contents

1. Features
2. Installation
3. Usage
4. Technologies Used
5. Authors
6. Machine Learning Explanation

Features ✨

* Apartment Filtering: Search and filter apartments for sale or rent in Tel Aviv, Jerusalem, and Haifa.

• Machine Learning Recommendations: Get personalized apartment recommendations based on your preferences.

• Interactive UI: User-friendly interface built with React for seamless navigation.

• Data-Driven Insights: Apartments data scraped and processed from madlan.co.il.

Installation ⚙️

Clone the repository and navigate to the project directory:

bash

git clone https://github.com/pazgu/Apartment_matcher.git
cd Apartment_matcher

Run the setup script to install all dependencies and start the application:

Python 3.11.X or greater is required

Before setting up the project, make sure you have the .env file with the MONGO_URI and the JWT_SECRET.

bash

./setup.sh

Note: Ensure you have npm, pip, and bash installed on your system.

Usage 🖥️

• Open your browser and navigate to http://localhost:3000.

• Explore apartments: Use the filter options to search for apartments to buy or rent.

• Get recommendations: Fill out the form to receive personalized apartment recommendations.

• Browse matches: Explore the top 20 apartment matches tailored to your preferences.

Technologies Used 🛠️

• Frontend: React

• Backend: Node.js, Express.js

• Database: MongoDB

• Machine Learning: Python, scikit-learn, pandas, NumPy

• Data Scraping: BeautifulSoup, requests

• Data Visualization: Jupyter Notebooks

• Algorithms: StandardScaler, KMeans, t-SNE, Euclidean distances

Authors 📝

• Paz Gueta - Backend developing using Node.js and MongoDB
• Steve Holof - Frontend developing using React
• Hanna Sofer - Frontend developing using React
• Yotam Zeevi Federman - Data scraping, Data preparing, Machine learning engineering

How does the Machine Learning model work ❓

The model uses clustering algorithms like KMeans and t-SNE to group similar apartments.

When you submit your preferences, it's treated as a "new apartment," the model finds the cluster the user's prefernces in, and using Euclidean distances it finds the 20 closest (most similar) apartments in the cluster to the user's preferences.