We are three students, currently enrolled in an IronHack data bootcamp.
Our project was to build a model that will predict the price of a house based on features provided in the dataset.
Within our python code we split the challenge into three components: ‘Explore the data, ‘Build the model’, ‘Visualise’.
Additionally, our task required using Tableau to graph the data and MySql to explore particular groupings and trends.
- Python 3.6
- MySQL 8.0
- Tableau 2020.4
Exploring the dataset we decided to change two variables:
- Utilize Uber’s hex_3 library rather zip codes.
- Create additional time variables from our date data: season, month_year, week_year.
Regression models used were:
- Linear Regression
- KNeighbors Regression
- Lasso Lars
- Polynomial
- SGDC Classsifier
- Random Forest Regression
- SGD Regressor
- XGB Regressor
- Gradient Boosting Regressor
Project has finished, and was done in 5 days. We found that Gradient Boosting Regressor, coupled with Power Transformer, produced the best model for house price prediction with the Washington dataset.
Useful websites that we used:
https://www.gislounge.com/modifiable-areal-unit-problem-gis/
https://towardsdatascience.com/stop-using-zip-codes-for-geospatial-analysis-ceacb6e80c38
https://hbr.org/2015/11/a-refresher-on-regression-analysis
https://www.upgrad.com/blog/types-of-regression-models-in-machine-learning/