Welcome to data mining project focused on predicting perfect rating scores for Airbnb listings. This project was completed as a group project for the course BUDT 758T: Data Mining and Predictive Analytics, using R as our primary programming language. As a group of passionate graduate students in MSIS of the University of Maryland College Park, we aimed to leverage advanced data mining techniques in R to analyze a comprehensive dataset and develop a predictive model that accurately predicts whether an Airbnb listing would achieve a perfect rating score or not. Extensive feature engineering was performed in R to enhance our model's performance.
Our project focused on predicting the likelihood of an Airbnb listing receiving a perfect rating score. By analyzing a comprehensive dataset, we aimed to determine whether a listing would achieve a 100% perfect rating score or not. The primary objective was to generate accurate binary predictions while minimizing false positives.
The dataset provided for analysis included various informative columns such as property access, accommodation details, amenities, availability metrics, cancellation policies, location information, pricing, and textual descriptions. These attributes offered valuable insights into the factors influencing perfect rating scores. The dataset can be accessed through the following Google Drive link: Dataset Link
We applied advanced data mining techniques and performed extensive feature engineering in R to enhance the performance of our predictive models. Through machine learning algorithms, we evaluated multiple models in R to identify the most effective one for predicting perfect rating scores. The Ranger model, implemented in R, demonstrated exceptional accuracy and precision.
Through charts and tables, we identified and analyzed 10 key variables that significantly influenced the perfect rating score. These variables provided valuable strategic insights into the factors contributing to perfect rating scores, helping hosts understand the important aspects to focus on.
Our predictive model, developed entirely in R, offers hosts the opportunity to optimize their listings and improve their chances of achieving a perfect rating score. By adjusting the identified variables and leveraging our feature engineering efforts in R, hosts can enhance the guest experience and attract more renters. Our findings and recommendations provide actionable insights to assist hosts in providing exceptional experiences on Airbnb.
The project report, along with the code, has been uploaded to this repository for easy access and reference. You can find the project report detailing our methodology, analysis, and results, along with the R code used for data preprocessing, feature engineering, and model development.
Finally, I would like to extend my gratitude to exceptional team members Aditi Patel, Meet Doshi and Vidit Vaywala for their invaluable contributions, collaborative spirit, and dedication throughout this project. I would also like to express my heartfelt appreciation to our professor Dr. Jessica Clark for his unwavering support and guidance.