Public Transportation Datalab
| Reference number | |
| Coordinator | RISE Research Institutes of Sweden AB - RISE SICS East |
| Funding from Vinnova | SEK 3 944 010 |
| Project duration | June 2019 - June 2021 |
| Status | Completed |
| Venture | Data-driven innovation |
| Call | Data lab and data factory as a national resource |
Purpose and goal
The main goal was to expand Trafiklab with tools for storing and publish historical public transport data. And to develop an environment and tools for data analysis, Big Data and AI. We have today collected real-time data (GTFS-rt) from 8 regional public transport authorities (PTA) and static data (GTFS, NeTEx) from 21 PTAs over a period of 1.5 years and we make this data available via a public API and a Python module. Possibility of quick access of structured data via a JupyterHub server is also available.
Expected results and effects
Today, there is a system that continuously downloads and stores the public transport data that is made available via Trafiklab. The goal of creating an environment for data analysis, Big Data and AI has been fulfilled by having a JuptyerHub server that has fast and direct access to the KoDa database and that we also expose a public API where the user can download historical data to their own computer for analysis. The tools above will make it easier for researchers, urban planners and other users to access and analyse public transport data independently of the affected traffic authority.
Planned approach and implementation
The data collection from Trafiklab´s open data APIs is done with Apache Nifi which is then stored, processed and transferred to an Apache Cassandra database. The historical raw data can be accessed via a public API and for quick access to the structured database there is access to a JupyterHub server. A Python module (pykoda) has been created and with associated examples, a new user can get a quick introduction to the Koda system. More information about APIs and how to get an account on the JupyterHub server can be read on the KoDa project website.