Just in time!

• 

  Notification for on-time arrival

Inspiration

Our biggest inspiration and starting point was to provide realtime, personalised service and alerts to users so they can get to their usual or preferred stop just in time. Our service does so by learning the behavioral patterns of the user (using historical data) and calculating the lead time to get to the stop based on this data and pattern.

This is a very valuable service for users using public transport in harsh finnish winters. Using our service they can be at their stop just in time so they don't freeze!

What it does

Send real-time notifications to iphone on the schedule of the user‘s usual bus and route. Suggests new route in case of disruption, delay or cancellation of usual bus. Advance reminder to leave (in minutes) of the usual or suggested new bus.

How we built it

Backend is written in python and we wanted to try swift for iOS apps. User's past data of destinations and arrival times for daily commute is stored on the server. Also the lead time required to leave is for now hard-coded to the server (in future this should be the learning algorithm). Using this data and combining it with the real-time transport data from HSL and user's current location from the phone, notifications to leave are sent to the app. For this user has to start the app themselves as currently push notifications are not in place for the prototype version.

Challenges we ran into

Real time HSL data is minimum and our app tends to fall back to scheduled data oftentimes. It was required to register to Apple developer program for 99$ to be able to send push notifications, so we had to skip the push notifications part for the prototype.

Accomplishments that we're proud of

Almost first time at Swift programming and Mika was pretty swift about it :) Also the fact that we are thinking bigger than just UI and focusing on the overall experience design and in a way also "machine-learning"

What we learned

Swift fundamentals, GraphQl, HSL routing and realtime API usage Service design and concept design keeping in mind persona of regular commuters in Finland and their user journey

What's next for Just in time!

Creating an algorithm to learn the commute patterns of the user (e.g. usual stops, usual schedules, number of stops and calculate precise lead time for the user to be at the stop based on their personal historical data. . The service has to continuously learn, adapt and build intelligence based on commuter’s behavioral pattern.

Built With

• hsl-realtime-api
• hsl-routing-api
• python
• swift