Team Name

Shadow Mantle

Team Members

Project Description

SchoolRoute

Smarter, Safer, and Future-Ready School Transport

Challenge: Optimising Transport Networks for School Kids

Team: Shadow Mantle

Project Type: Web Application

Focus: Student experience, public transport accessibility, data-driven planning, and future transport innovation

The Problem

Public school transport systems are often outdated, inflexible, and difficult to navigate—especially for students and working families. Key issues include:

• Inconsistent bus services that are often late, overcrowded, or poorly aligned with school schedules.
  
• Congestion around schools due to high reliance on parent drop-offs.
  
• Lack of accessible tools for students and families to plan their journeys.
  
• Limited data-driven planning for future transport needs, especially in growing suburbs.

These challenges disproportionately affect students in outer suburbs, low-income areas, and those with limited access to private transport.

Our Solution: SchoolRoute

SchoolRoute is a student-first, data-driven web application that empowers students, families, and planners with intelligent insights into school transport. It provides real-time and historical data on bus route reliability, accessibility, and usage—making school travel safer, more predictable, and more efficient.

Implementation of an AI chatbot provides user-friendly, succinct information that is easily digestible for students, which prioritises the safety and reliability of different bus routes to aid student decision making.

SchoolRoute Features

• Smart Route Finder
  Students and parents can enter their address and desired arrival time to receive tailored school bus route suggestions.
  
• Reliability Analyzer
  Our AI model shows how consistent a route is—highlighting whether it tends to be early, late, or on time.
  
• Crowdedness Estimator
  Using student distance data and route mapping, we estimate how busy a route is likely to be during specific time windows.
  
• Planning Feedback Loop
  Route performance data is packaged for schools and transport authorities to identify gaps, inefficiencies, and areas for improvement.

Future Vision: The Real Innovation

While SchoolRoute focuses on reliability and accessibility, SchoolRoute’s long-term vision is to become a comprehensive AI-powered transport planning and navigation platform for school communities. This is where our most transformative impact lies.

1. Generative AI for Predictive Transport Planning

Use generative models (e.g., OpenAI API) to simulate future transport scenarios based on:

• Population growth (using census and planning data)
  
• School expansion or rezoning
  
• Urban development and infrastructure changes

This will provide valuable insight for planning authorities by helping them find possible avenues for reducing congestion in the future by generating potential new bus route proposals that adapt to projected needs.

General outline of process:

• Forecasting student demand by area: Joining population projections (SA2) to school enrolment trends and allocate students to likely schools via catchments
  
• Generating time-of-day trip profiles using school bell times, adjusting for observed patronage seasonality
  
• Building a multi-layer graph of transit layer and pedestrian/road layer
  
• Estimate loads on current network: assign forecast O–D demand to existing services via time-dependent shortest paths; compute stop/segment load and reliability
  
• Create candidate new routes: k‑shortest paths between demand centroids and schools with filter capabilities
  
• Optimise a route set & frequencies to maximise covered student demand and minimise generalised travel cost, subject to fleet, headways, ride time, and safety
  
• Validate the loads and predictions align with observed patterns

Datasets to be used:

• GTFS for Transport Canberra (routes, trips, stop times, shapes, frequencies).
  
• ACT Bus Routes (shapes) for spatial ground-truth and for sanity checks against GTFS.
  
• Timing Point Locations, anchor points used for punctuality/OTR metric. Open data portal dataACT
  
• Daily Public Transport Passenger Journeys (by service type)—to calibrate demand seasonality and school‑day uplift.
  
• OpenStreetMap road/walk/cycle network (to compute access/egress and “safe path” features).
  
• ACT Schools Census (enrolments, sectors, locations)—time series by school and year
  
• ACT Population Projections (SA2/SA3 + age/sex) - 2022–2060; school‑age cohorts to project future demand by area
  
• ASGS geographies (Mesh Blocks/SA1/SA2) - for consistent spatial aggregation and school catchment approximation

The team were not able to complete this in time for submission

2. Graph Analytics for Network Optimisation

• Model the entire school transport system as a graph of nodes (bus stops, schools, Park & Ride locations) and edges (routes).
  
• Apply graph algorithms to:
  
  • Identify underserved areas and transport deserts
    
  • Optimise route efficiency and coverage
    
  • Minimise student travel time and transfer complexity
    
• Enable dynamic re-routing based on real-time data (e.g., traffic, delays, weather).

Graph Construction:

• Build a multi-layered graph where:
  
• Nodes represent bus stops, schools, Park & Ride locations, and key pedestrian access points.
  
• Edges represent bus routes, walking paths, and road segments with attributes like travel time, reliability, and safety.

Datasets to be used:

• GTFS for Transport Canberra (routes, trips, stop times, shapes, frequencies)
  
• ACT Bus Routes (shapes) for spatial validation
  
• OpenStreetMap (road/walk/cycle network) for pedestrian and access path modelling
  
• ACT Schools Census (locations, enrolments)
  
• ACT Population Projections (SA2/SA3) for future demand estimation
  
• ASGS geographies for spatial aggregation and catchment modelling
  
• Real-time traffic and weather feeds (for dynamic routing simulation)

3. Student Sentiment & Safety Feedback Engine

• Collect structured and unstructured feedback from students on:
  
  • Route safety
    
  • Cleanliness
    
  • Comfort
    
  • Driver behavior
    
• Use natural language processing (NLP) to extract themes and trends from open-text feedback.
  
• This valuable analysis can be used by schools and transport authorities to prioritise improvements.
  
  • Empowers students with a voice in their daily commute.
    
  • Gives schools and transport providers real-time insights into safety and quality issues.
    
  • Builds trust with parents through transparent improvement tracking.
    
  • Reduces risks by prioritizing safety-driven interventions.
    
• Will use user input and process this data into standardised data for the model to use

Impact

For Students & Families:

• Reduces stress and uncertainty around school travel
  
• Promotes independence and confidence in using public transport
  
• Encourages sustainable travel habits

For Schools & Planners:

• Provides actionable insights into route performance and demand
  
• Supports data-driven decision-making for future infrastructure
  
• Helps reduce congestion and improve safety around schools

For Communities:

• Reduces environmental impact through smarter transport planning
  
• Builds a foundation for future-ready, AI-enhanced public services

Data Story

Data Story

Use of data within the application:

This is provided as a data flow diagram:

Dataset information:

1. Australian School List from Peclet Technology

This publicly available dataset provided us with detailed information for each school in the ACT, including the school name, type, sector, state, postcode and location.

This data was filtered to only include schools within the ACT and plotted on a map using the provided coordinates. This comprehensive dataset also provided us with the required search completions for our web application, allowing users to quickly find their school.

Additionally, the data of the school selected by the user was used as context for the AI model, helping it assess the route more effectively.

2. ACT School Bus Services

This government dataset contains a list of all the school bus services in ACT, which we organised in a table on our website. We also implemented search capability, making it an easily accessible way of looking up service data effectively.

This data did not require preprocessing, as it was downloaded in the form of a fully intact csv file that we could easily map to a table and display.

3. ACT Public Bus Routes

This data was essential for us, as it contained an array of longitude and latitude coordinates for every public bus route in the ACT. After processing the data and converting it into appropriate values, we could plot these points as polylines on a map to display the bus routes.

Like the school list dataset, these coordinates were also provided to the AI model as context to effectively assess accessibility by calculating the distance.