Back to Projects

Team Name:

Tiny Happy People Hacking


Team Members:


Evidence of Work

T-Chat

Project Info

Tiny Happy People Hacking thumbnail

Team Name


Tiny Happy People Hacking


Team Members


3 members with unpublished profiles.

Project Description


For GovHack 2019 we’ve taken on the challenge of improving public transport experiences into the future with open data.

We've created a real-time location-based chat platform for public transport users, T-Chat.

Problem

If you travel on public transport you’ll be aware everyone travels in their own little bubble.

They get on, don’t look or talk to anyone and settle down with their phone, tablet, laptop or, in rare cases, a book or newspaper. There’s no sense of community, even where people have been travelling on the same line at the same time for years.

Plus there’s few ways to access information about areas folks are passing through or to interact in engaging ways with fellow travellers.

Our Solution - T-Chat

T-Chat is a location-based conversation, information and entertainment app that leverages real-time transport open data to connect you to the community inside your vehicle.

T-Chat allows folks on the same public transport vehicle to form and sustain mobile communities, interacting with each other and with an AI, ChatDriver.

For GovHack we have created T-Chat as a web-accessible chat system, which would be integrated into an app framework in future sprints.

We integrated T-Chat with ACT Light Rail and Queensland train, ferry and bus real-time location data to connect anyone using the app to a custom chatroom created for each vehicle based on vehicle and user real-time GPS location. We were unable to get similar NSW or ACT real-time bus data working within our app in the available time.

The ChatDriver AI is only very basic in our GovHack due to time and selecting a learning AI rather than a pure scripted chatbot to allow it to expand to support a wider variety of requests and languages over time. This capability would be expanded in the next sprint to provide information about travel routes and arrivals time, the vehicle being travelling in (make, model, history, driver experience, etc) and provide the capability to ask simple travel support questions and comments about individual travel experiences.

This latter data can be used to continue to improve the public transport network.

ChatDriver would also become a more effective tour guide in future iterations. While we began integrating open data on heritage locations and landmarks, we did not fully activate this in the app due to a lack of tour information in these open datasets.

Our plans would include expanding on heritage and landmark listing data and more precisely using the vehicle’s real-time GPS location, such that ChatDriver can point out landmarks along the route in a timely manner (given the speed and direction of travel), helping residents connect more deeply to their city and assisting tourists to find locations they would otherwise have missed.

We also have aspirations to expand ChatDriver to support quiz-based gaming, allowing folks to compete on their knowledge of a city within a vehicle, or to form teams that compete against routes as part of a city-wide leaderboard. This would also leverage open data to ask questions related to the city.

Again this capability is easy to add within our framework, simply requiring additional time to implement.

Finally, we planned on using open data to enable access to useful and interesting information such as how much petrol and carbon dioxide a T-Chat user, or all commuters on a vehicle, saved by choosing public transport. We are yet to add this functionality, leveraging NSW and other open datasets, as we found they required too much additional data cleaning to provide an accurate result within the timeframe available.

We had hoped to develop T-Chat and ChatDriver further in the time we had, however challenges with many of the datasets not being fit for this purpose mean that this initial version of the service is more of a sketch of what T-Chat could become with appropriate data and support.


Data Story


For T-Chat we started with the base real-time and movement and standard data for public transport, using Real-Time GTFS Feed from the ACT (light rail) and South East Queensland's General transit feed specification (GTFS) (buses, trains & ferries) using OpenAPI - which we were able to integrate rapidly and gave us core information such as stops, distance from stops and traffic levels. We were not able to integrate the ACT buses due to their requirement for registration ahead of time.

From here we build a location-sensitive chat service which detected a user's location relative to a public transport vehicle and created custom chat windows for each vehicle, with information on their location as they move.

We created a simple AI that could leverage landmark data, map it using Open Street Maps against the location of the vehicle and respond to enquiries as a basic travel guide - though we have not integrated view distances at this time.


Evidence of Work

Video

Homepage

Project Image

Team DataSets

Real-Time GTFS Feed (Queensland)

Description of Use Collection of all Queensland active public transport vehicles to match against for joining local chat rooms.

Data Set

Real-Time GTFS Feed (ACT)

Description of Use Used to join to local chat rooms based on the users location data blended with the vehicle data

Data Set

Standard GTFS Feed (ACT)

Description of Use Collection of all ACT active public transport vehicles to match against for identifying the stations and stops.

Data Set

Open Street Map Locations

Description of Use Used to identify landmarks in the Canberra region

Data Set

Challenge Entries

Public Transport for the Future

How might we combine data with modern technologies - such as AI/ML, IoT, Analytics or Natural Language interfaces - to better our public transport services. Outcomes could take the form of new commuter experiences, reduced environmental impact, or helping plan for the future.

Eligibility: Use any open dataset to support your entry.

Go to Challenge | 45 teams have entered this challenge.

🌟 What's the coolest way to travel across the city?

Using datasets which map urban heat and green cover across Greater Sydney, we challenge you to develop a tool which visualises green routes through the city. Help people avoid urban heat and move across the city in comfort by mapping out green streets and pathways which connect shopping centres, public transport stops and public spaces.

Eligibility: Must use at least the Urban Heat & Green Cover dataset(s) from the SEED portal.

Go to Challenge | 18 teams have entered this challenge.

Reducing CBD Traffic Congestion

How to reduce traffic congestion or parking problems in CBD?

Eligibility: Use any open dataset to support your entry.

Go to Challenge | 39 teams have entered this challenge.

🌟 Canberra 2029 – Inclusive; Progressive; Connected

How do we use data from the past to predict a better future for Canberra? How do we best support the diversity of our community? Optimise the way we travel and transport goods throughout our city? Predict the jobs of the future – and the skills needed for them? Connect our citizens with their environment?

Eligibility: Must use at least ONE relevant/related dataset from www.data.act.gov.au

Go to Challenge | 21 teams have entered this challenge.

Queensland OpenAPI

Create a project using one or more of Queensland's Open-API’s

Eligibility: Your solution to this challenge should access at least one Queensland Open Data API. You can mix and match this with other API’s or data

Go to Challenge | 39 teams have entered this challenge.

Training AI models to deliver better human outcomes

For an outcome create two AI models based on contrasting incentive systems and examine the differing impacts on the defined outcome.

Eligibility: Must make use of any open dataset and apply two different incentive systems

Go to Challenge | 12 teams have entered this challenge.