2024
Buenos Aires real estate, exploratory analysis
EDA of 146,660 property listings to help appraisers value homes with less subjectivity.
Problem
Property appraisal is slow and subjective.
Data
Property listings published in 2019–2020 (the Properati dataset used in the course). The CSV is not included in the repo.
Method
Cleaning, IQR-based outlier filtering on area and price, and correlations by property type.
Result
For the 3 most-listed property types in Buenos Aires City (91,485 listings), rooms and bedrooms are the best individual price predictors, with moderate correlation.
Stack
- Python
- pandas
- Seaborn
- Matplotlib
Context
Course: Professional Practicum I · 2024