Aqar.com: Price Prediction of Rent Apartments in Riyadh

Hatim Alshehri

Abstract:

Aqar is a website specialized in Saudi real estate. The project's goal is to extracting data, for rent apartments in the Riyadh region from Aqar website, and generate a regression model to predict the prices of apartments.

Data:

Data scraped from Aqar website using Selenium tool.

The data scraped is described by 14 features as follows:

• Field Description:

Field Name	Description
District	Apartment districts/neighborhoods
Category	(e.g., snigal/fmaily)
Bedrooms	Number of bedrooms
Livingrooms	Number of Livingrooms
Bathrooms	Number of Bathrooms
Furnished	Does apartment has Furnished (e.g., yes/no)
Kitchen	Does apartment has Kitchen (e.g., yes/no)
Garage	Does apartment has Kitchen (e.g., yes/no)
Elevator	Does apartment has Elevator (e.g., yes/no)
AC	Does apartment has AC (e.g., yes/no)
Region	Apartment region in Riyadh (e.g., west/north)
floor_number	Apartment floor number
AGE	Property age
Price	Apartment rent price per year

Design:

A regression model analysis was conducted that encompasses many features and among them the apartment's price. To gather apartment data Aqar.com has been scraped, one of the top and most visited online real estate agencies in KSA. I utilized several regression models and tested for the best fit; to ensure the best predictor tool.

Algorithms:

• Clean: The dataset had duplicate observations, NaN values, and spaces in between the categorical features, so we used pandas library to prepare the data for the regression model.

• Preprocessing:Used transformation methods in order to apply to standardize the values at an equivalent scale and to linearize some of the features that are not linear.

Models

Tools:

Language: Python:

• Data Scraping libraries: Selenium

• EDA Libraries: Pandas, numpy, seaborn, matplotlib، Missingno

• Model Building Libraries: sklearn and Model Testing libraries sklearn