Predicting real estate prices from urban vitality
A real estate appraisal is multifaceted due to numerous property price variables. The factors known to influence the price are location, structure, and neighborhood. However, these features are property-centric; it does not consider the built environment's impact. Jane Jacobs introduces the the...
Saved in:
| Main Author: | |
|---|---|
| Format: | text |
| Language: | English |
| Published: |
Animo Repository
2022
|
| Subjects: | |
| Online Access: | https://animorepository.dlsu.edu.ph/etdm_softtech/3 https://animorepository.dlsu.edu.ph/cgi/viewcontent.cgi?article=1002&context=etdm_softtech |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | De La Salle University |
| Language: | English |
| Summary: | A real estate appraisal is multifaceted due to numerous property price variables. The factors known to influence the price are location, structure, and neighborhood. However, these features are property-centric; it does not consider the built environment's impact. Jane Jacobs introduces the theory of urban vitality as the conditions cities should exhibit to ensure a livable built environment. Although some characteristics in the existing literature overlap with urban vitality, many remain unexamined. Only a handful of research has investigated the influence of urban vitality on real estate prices.
In this work, we use multiple data sources to develop an XGBoost model that predicts real estate prices and identifies the features influencing house prices in the Philippines. The final model performed an R2, MAE, and RMSE scores of 0.70798, 0.001412, and 0.013449, on the test sets. The model can also achieve the R2, MAE, and RMSE scores of 0.68734, 0.001404, and 0.013702, on the holdout sets. The structural features contributing to the price are land size, floor area, number of bedrooms and bathrooms, property class, and transaction type. The walking distance scores for typical consumer destinations were not among the top essential features. The model considered the urban vitality features, block area, and closeness to daily places as important contributors.
We also discovered essential features that previous real estate modeling literature did not encounter —internet connectivity, sports facilities, fire systems, receiving facilities, and security systems.
Lastly, we develop a decision support tool to visualize the features influencing the property's price.
Keywords: urban vitality, real estate prediction, open data, feature database, xgboost, data visualization |
|---|