SmartBikes: Optimizing Bike-Sharing Systems for Urban Mobility

Guide to our Repository:

Data Preparation:

• demand_score.ipynb: to categorise demand levels for each station, outputs station_with_demand.csv
• distance_matrix.ipynb: recalculates distances between stations in US metric feet, outputs distance_matrix.csv
• station_with_demand.csv: a cleaned dataset of Manhattan trips, with each station classified into a demand level based on binning

First Stage:

• bikeprice.ipynb: to determine the unit value assigned to a bike given a station's trip data and popularity
• first_stage.py: used to draft and test the first stage model
• helpers.py: helper functions used for data input preparation for the first stage model
• first_stage_func.py: actual first stage function used for the final simulation of our model
• This stage was dependent on:
  1. station_with_demand.csv

Second Stage:

• avg_ride_count_per_week_weather.csv: a csv generated from weather_demand.ipynb. It contains the average ride count per week in autumn based off of various weather conditions (normal, sunny, cloudy, rainy)
• daily_ride_count_by_weather.csv: a csv generated from weather_demand.ipynb. It contains the average ride count daily in autumn based off of various weather conditions (normal, sunny, cloudy, rainy)
• weather_demand.ipynb: a Jupyter notebook that wrangles the 2023 citibike tripdata and 2023 NYC weather data to classify bike trips in Manhattan based off on weather conditions
• second_stage_v2.ipynb: used to draft and test the second stage model
• second_stage_func.py: actual second stage function used for the final simulation of our model
• This stage was dependent on:
  1. distance_matrix.csv
  2. station_with_demand.csv

Model Simulation:

• simulation.ipynb: full simulation file This is dependent on:
  1. first_stage_func.py: function format for first_stage.py
  2. second_stage_func.py: function format for second_stage_v2.ipynb

All other files were used for intermediary steps or drafting to derive values for our analysis.