The global car rental industry is expected to reach an estimated $120 billion by 2025 with a CAGR of 6.1% from 2020 to 2025.
However, after a long search, i was unable to access an open-source project related to it, and the data is also scarce.
Finally, I found a database file consisted of a 330MB JSON file with a heavily nested format., that was extracted from the (TURO) website. consisted of 36000 car rental events.
therefore I decided to make this project open source. I tried to explain step by step starting from reading the database file and converting it to flatten data through the EDA included the spatial analysis as well.
and comparing between -Random Forest and XGBoost- to predict daily rent rates.
It's also includes how to deal with different data sources as I have added a population dataset.
I did my best to explain the steps of hyperparameter tuning in details for beginners as well .
Packages used :
data visualization :
Plotly.
Seaborn.
Matplotlib.
ggplot.
Spatial analysis:
Geoviews .
Folium .
go.figure plotly.
geopandas.
hyperparameter tuning:
GridSearch.
Randomsearch.
manually.
To download data here