Dubai Rent Prediction System: Data-Driven Intelligent Analysis for Real Estate Investment Using 260,000 Real Data Entries

Based on 263,000 registered rental contracts from the Dubai Land Department, this open-source project builds a machine learning rent prediction model with an accuracy rate of 95.10%, and provides investors with real-time market insights and decision support through an interactive Power BI dashboard. The project is maintained by shahrouribader-bit, and the source code is available on GitHub (link: https://github.com/shahrouribader-bit/BI-Rental-price-prediction-Dubai). It was released on June 2, 2026.

Why Choose Dubai?

As a global financial and tourism hub, Dubai's real estate market features high data transparency, strong activity, high internationalization, and standardized regulation (mandatory registration of rental contracts).

Data Scale and Quality

The project uses 263,000 DLD-registered rental contracts, which have statistical significance, market representativeness, time span coverage, and authenticity guarantees, avoiding interference from fake listings.

Data Collection and Integration

Data sources include DLD official rental contracts, property features, geographic location, and macro market data.

Data Cleaning and Preprocessing

Covers data type conversion (unified units, standardized formats), missing value handling (imputation strategies), data merging and aggregation (multi-source integration, derived features such as rent per unit area).

EDA and Feature Engineering

Through visual analysis of price distribution, correlation, and time series trends; constructs numerical (area, rent per unit), categorical (property type, region), time (season, contract term), and geographic (latitude/longitude, distance to business districts) features.

Machine Learning Models

Uses algorithms such as linear regression, ridge regression, random forest, and XGBoost. Evaluation metrics include R², RMSE, MAE, MAPE, achieving an accuracy rate of 95.10%.

Power BI Dashboard

Provides market overview (rent trends, heatmaps), investment analysis (return rate calculation, regional comparison), property screening (multi-dimensional filtering, anomaly detection), and trend prediction functions.

• Individual Investors: Evaluate property rental potential, pricing reference, regional comparison, and timing judgment.
• Real Estate Agents: Precise pricing recommendations, customer consultation support, and professional market report generation.
• Developers/Asset Management Companies: Product positioning, rent strategy optimization, and portfolio monitoring.
• Academic Researchers: Market rule analysis, ML application research, and data science teaching cases.

1. Real Data Scale: 263,000 real rental contracts, close to industrial application level.
2. End-to-End Solution: Covers the complete process from data collection, cleaning, analysis, modeling to visualization.
3. Business-Technology Integration: Solves investors' decision-making pain points and transforms technology into practical value.
4. Interpretability and Practicality: The Power BI dashboard makes complex models easy to understand and use.

Data Acquisition

• Seek government open data, obtain desensitized data through platform cooperation, use compliant web crawlers, or purchase data.

Technology Selection

• Python (pandas/sklearn) for data processing and modeling; Power BI/Tableau for visualization; Flask/Streamlit optional for web applications.

Model Optimization

• Try deep learning and time series models; introduce external data such as economic indicators; build real-time prediction systems or API services.

• Data Limitations: Only covers the Dubai market; models need regular updates; some features may be missing.
• Model Limitations: Accuracy may fluctuate in practical applications; extreme market conditions (e.g., pandemics) are beyond the training scope, and black swan events cannot be predicted.
• Application Limitations: Prediction results are for reference only and do not constitute investment advice; decisions need to consider factors such as law and taxation.

The Dubai rent prediction project demonstrates the potential of big data and machine learning in the PropTech field, providing investors with data-driven decision-making tools. The core value of data science lies in extracting actionable insights from massive data, and this project is an excellent model for learning and practice. Whether you are a data scientist, real estate practitioner, or investor, it is worth in-depth research and reference.

"In the data-driven era, intuition and experience are still important, but only when they dance with data can the most beautiful investment melody be played."