What is our transport optimisation offering about?

• AI tools for bus congestion optimization

• Data lake capabilities

• Passenger counting sensors, along with:

  • Mobility analysis

  • Bus vehicle/Bus stop occupancy prediction models

  • Passenger waiting times/returning visitors analysis

• Mixture models – to identify factors most responsible for targets (e.g. passenger counts)

• Other fine-tuned models‍

Business problems addressed

Collecting too much data, leaving it unused → companies have multiple data sources with varying structures, which would be much easier to manage in a data lake

Lack of information → many transport companies do not know the exact congestion, mobility, or passenger-related data

Lack of efficiency (late buses, crowded buses, etc.) → with optimized schedules, companies can avoid congestion and delays, leading to better use of their resources. New mobility data and other statistics can also help to see useful patterns, leading to more optimal planning

Old (or no) data collection methods → some companies still use manual data collection which could be easily automated with sensor data or available data source integration‍

How does it work?

• The client provides available data sources and their descriptions (to be used for the models/tools), provides exact sensor deployment locations, permissions, ensure powering capabilities (if sensors are chosen)

• The client generates more datasets upon request

• Provides additional rules for the models (e.g. ‘congestion optimization model cannot disrupt bus driver breaks’, etc.)

• Provides authorisations, to sign agreements regarding GDPR, etc.

Results

We create a system with the desired solution (congestion optimizing, data analysis/statistics, data lake, mixture models), containing:

• A web application

• Deployed sensors (if applicable)

• API endpoints, along with all used Python scripts

• Technical and user documentation

This enables the client with:

• Additional data collection

• Data source aggregation

• Data analysis (mobility, congestion, waiting times/returning visitors)

• Automatisation of manual data collection

• Reduction in human error and needed resources

Project phases & timeline

Our services are fast and easy to implement, without research uncertainties.

1. Free consultation & recommendations → Within 2 weeks of initial contact

2. Analysis of processes & roadmap what to do → Within 2 weeks of signing the contract

3. Model training and integration → Within 4 weeks of the day when the client provides the data

4. New functionalities, retraining → Upon separate agreements/invoicing

Overall duration from around 3 months (Congestion tools, Mixture models) to 6-12 months (Sensor-based solutions).