Real_Estate_Project_Fatma

Objectives

The primary objective of this project is to collect, clean, and analyze rental property data from various websites. The goal is to create a comprehensive dataset suitable for exploratory analysis and predictive modeling related to property rentals in Oman.

Websites Used

• Bayut: https://www.bayut.com
• Osooq: https://www.osooq.com

Data Collection

• Used web scraping techniques to extract property listings from Bayut and Osooq.
• Employed requests and BeautifulSoup libraries to fetch and parse HTML content.

Data Cleaning

• Loaded the scraped data into Pandas DataFrames.
• Merged datasets from both sources into a unified DataFrame.
• Handled missing values using appropriate strategies.

Size Cleaning

• Converted size descriptions into numerical format to enable quantitative analysis.

Governorate Extraction

• Parsed location strings to extract governorate.

Price Normalization

• Cleaned and standardized price data to ensure consistent numerical formatting.

Additional Features

• Created a new column for price per square meter.
• Encoded categorical features using LabelEncoder for compatibility with modeling algorithms.

Modeling Approach

• Implemented a Decision Tree Regressor to predict rental prices based on engineered features.
• Split the dataset into training and testing sets to evaluate model performance.
• Assessed accuracy using:
  • Mean Squared Error (MSE)
  • R-squared (R²)