Objective : The goal of this project is to gain insights on important factors that influence the rental values of homes and apartments in Phoenix and Tampa.
Measurement of success : The performance of our model will be tested using mean squared error.
Questions to answer :
- What are the similarities and differences in the rental markets in Phoenix and Tampa?
- How accurately can the model from one city predict rental values in the other?
- How important are historical rental values to predict future ones?
- What factors drive the prediction of rental values in each market?
- What actionable insights can be derived from our analysis? For example, can we increase the rental prices in certain zipcodes if historical data tends to underestimate the rental value.
Data Sources
- Zillow Historical ZRI Data Multi Family Homes Historical ZRI data from 10/2010 - 1/2021
- Google Bigquery Data from geo_us_boundaries Data from ACS
Data Aggregation : By zipcodes for cities Phoenix and Tampa
Target variable : Zillow Rental Index (ZRI)
Tools : Google BigQuery, Jupyter Notebook
Analysis/Machine Learning
- Missingness/Imputation,
- Feature Selection,
- Feature Engineering,
- Ridge/Lasso regression,
- Tree ensembles,
- Boosting,
- Statistical Analysis,
- Residual Analysis
Our Team
- Chitra Sharathchandra: Software developer
- Gabriela Huelgas Morales: PhD in biomedical sciences
- Juan Vasquez: Operations professional
- Yukti Kathuria: MS in Aerospace Engineering
- Guillermo Ruiz: MS in Economics
Folder breakdown in repository
- Chitra: SQL Exercise rough code, Phoenix feature extraction and analysis, EDA, Cluster analysis
- Gabriela: SQL Exercise rough code, Modeling nationwide ACS/ZRI data, Phoenix EDA and residual modeling
- Juan: EDA on ACS zipcode data, some preliminary SQL test code, Phoenix Rent growth rate
- Yukti: EDA on ACS zipcode data, some preliminary SQL test code, Feature engineering for Phoenix,
- Guille: EDA code on ACS zipcode data, Modeling and Statistical Analysis for Phoenix
- SQL_exercises: Final code for SQL task