baywheels_lyft_analysis

Analysis of the baywheels data found and scraped from here: https://s3.amazonaws.com/baywheels-data/index.html

Goals of this project

Exploration

Exploration of the data to see what's possible

Data cleaning/wrangling

Creating new features and wrangling data to create new analytical possibilities (ex: interpolating the cost of a ride based on website pricing information).

Machine learning

Predicting the revenue of a ride or stop by gender and/or age, time, etc.

Algorithms

Determining the best place to position the bikes/stops for maximum revenue

Setup

You will need Python 3.7 and pip, the package manager. Please run a virtual environment and install the dependencies listed below:

Necessary packages

Please run a virtual environment and install the dependencies listed below:

1. pandas
2. bs4 (BeautifulSoup)
3. Jupyter Notebook

The other packages are native to Python 3.x and should work with the import statements in the code.

Layout and file/folder structures

All baywheels csv's are scraped from the link at the top of this readme.

These data are scraped and compiled in the baywheels-data folder (not included for repo size issues).

To access the data:

1. Run through scraping to get all the CSVs.ipynb and save them to your local directory of choice.
2. From there you can run master_df_cleaning.ipynb if you have lots of RAM, otherwise try dev_df_data_cleaning.ipynb. Note: make sure to update the directory variables to the appropriate path on your computer!

Note: Go to Full_Pipeline.ipynb to see all the code from data cleaning to EDA to statistical tests.

This notebook requires the full dataset which was not uploaded to the repo. You can reconstruct it by using the master_df_cleaning
notebook mentioned above. Alternatively, if you don't want to do any of that or run cells, you can also view Full Pipeline Preview (No Coding Required).pdf.

The code in here is intended to be standalone.