Modeling Toronto bike-share dataset using k-means clustering
Toronto Bikeshare Data
Toronto Bikeshare Data on Kaggle
https://www.kaggle.com/jackywang529/toronto-bikeshare-data/download
The Bike Share Toronto Ridership data contains anonymized trip data, including: Trip start day and time, Trip end day and time, Trip duration, Trip start station, Trip end station, User type
In this dataset, we have the bike sharing information form 2017 and 2018.
Usage Information License Other (specified in description)
Visibility visibility Public
Provenance Sources (https://open.toronto.ca/dataset/bike-share-toronto-ridership-data/)
Collection methodology Collected from https://open.toronto.ca/dataset/bike-share-toronto-ridership-data/ on 2/4/2020.
Maintainers Dataset owner JackyWang (https://www.kaggle.com/loganalive)
Updates Expected update frequency Annually
Last updated 2020-02-05
Date created 2020-02-05
Current version Version 2
- trip_id: unique identifier for each trip
- trip_start_time: time when the trip started
- trip_stop_time: time when the trip ended
- trip_duration_seconds: duration of the trip measured in seconds
- from_station_id: station ID where the trip started
- from_station_name: name of the station where the trip started
- to_station_id: station ID where the trip ended
- to_station_name: name of the station where the trip ended
- user_type: identify if the user has a membership or purchased a pass
This dataset is from Toronto Parking Authority, published on https://open.toronto.ca/dataset/bike-share-toronto-ridership-data/. You may find the latest complete dataset from there.
The data is licensed under: Open Government License - Toronto
It might seem to be intuitive that bikeshare users would travel in the reverse direction in the evening than in the morning.
But how true is that? If that's the case, would it reduce the value of convenience of bikeshares being available everywhere, and should they just get their own bikes? Do we also need a crew to pick up and restore all the bikes to initial distributions across the city before the next morning?
- Find out what percentage of bikes are returned to its initial location?
- How many bikeshare trips usually visit more than just start/end points, but visited some other points before the end of the day?
References for code programming/articles
- Memahami K-Mean Clustering Pada Machine Learning Dengan Phyton
- Exploring Toronto Bike Share Ridership using Python
- Bike availability in the station
- See the flow of bikes
- Find out the Top 10 Stations with the highest number of a two-way traveler using EDA?
- How to predict the flow of bikes. Are bikes returned to their initial location? Or are there flows of bikes from one region to another?
Top 10 Stations with the highest number of two-way traveler
from_station_name to_station_name
Bay St / Queens Quay W (Ferry Terminal) 3515
Humber Bay Shores Park West 2162
York St / Queens Quay W 1969
Ontario Place Blvd / Lakeshore Blvd W 1746
Tommy Thompson Park (Leslie Street Spit) 1742
HTO Park (Queens Quay W) 1606
Lakeshore Blvd W / Ellis Ave 1562
Lakeshore Blvd W / Windermere Ave 1238
Lakeshore Blvd W / Ontario Dr 1201
Dockside Dr / Queens Quay E (Sugar Beach) 1028
Methode to Clustering Model
- KMeans on gps coordinates (nbr_cluster = 20) (Clustering of the stations in 20 regions)
- Mapping Station. Coordinates of cluster centers and map each station with its region
- Plotting Cluster
- Sum of all travel between 2 nodes during a period
- Extract flows and creates graphs flows by time, stations, and regions.
Plotting clustering of the stations in 20 regions
Plot the graphs for stations for weekdays in 2018
Plot the graphs for regions for weekends in 2018
