Big Data Innovation Team
The Big Data Innovation team was created in 2015 with the mission of leveraging emerging transportation datasets together with existing City data to develop a new understanding of transportation issues across all modes in the City. The Big Data Innovation Team conducts practical analyses of transportation data to be able to more easily measure the impact and benefits of policies and solutions. The Team partners with researchers and will focus on providing excellence in the communication and visualization of findings. The team’s code is Open Source and available in repositories prefixed by “bdit” on the City of Toronto’s GitHub Page.
The Innovation Team will begin by building on some current work being conducted by Transportation Services, including:
- collaborative effort with the University of Toronto to estimate traffic volumes on all segments of the City at all times, informing the future development of a comprehensive multimodal count strategy
- partnering with Ryerson University to analyze historical travel data on city expressways and streets and develop strategies for evaluating projects and interventions
- collecting data to monitor performance of the King Street Transit Pilot Overview
- developing a Big Data Strategy and Work Plan for Transportation Services to determine ways to make this type of information available to map out how the team will proceed
- vetting products and services that might be useful in assisting the City in better decision-making and investments
The City of Toronto commissioned a study from the McMaster Institute for Transport & Logistics (MITL) that used INRIX GPS probe data to analyze historical traffic data on City expressways and arterials. The study looked at three time periods (August-December 2011, July – December 2013 and January- December 2014).
The study is the City’s first foray into making practical use of this type of probe vehicle data. Much of the work done here is on the cutting edge, and through this study the City of Toronto and MITL are developing methodologies that will evolve and improve over time, allowing the City to track and monitor congestion levels across the City, year by year. The City recognizes that the methodology may not be perfect or final, but it is a strong first step in what will be an on-going process. While there are varying opinions on the how congestion is defined, what is most useful is that a strong academically supported methodology has been developed that can measure changes in congestion over time.
The overall study was conducted in three parts:
The first phase identified that the single most congested days occurred on days during which there was snow or rain. While this is in many ways expected, these results illustrate the role of weather in travel conditions and demonstrate the utility of these approaches when analyzing Big Data for performance monitoring.
City Congestion Trends
The second phase estimated changes in traffic congestion over the three year period from 2011 to 2014 by looking at annual, monthly, daily and hourly variations in performance metrics, including speed, delay, and unreliability. The study found that congestion did materially grow from 2011 to 2014, but the growth was uneven and congestion was in fact lower in 2013.
Corridor Report Cards
The final phase included a set of corridor report cards for 36 corridors across the City. Corridor report cards provided comparable snapshots of changes in performance between 2011 and 2014, hourly speed profiles for typical days of the week, and measures of unreliability. Results identified uneven changes in congestion over time among City roadways and expressways.
Transportation Services partnered with Evergreen CityWorks to host TrafficJam, a 48-hour traffic data Hackathon at the Evergreen Brickworks from October 2nd to 4th, 2015. At TrafficJam, 135 participants collaborated with traffic analysts, government officials and data collectors to offer insight and solutions toward better understanding and management of transportation issues. New ideas were brainstormed, a range of large datasets were analyzed and tools were built to provide for better decision making. At the TrafficJam Expo the most promising entries were recognized and awarded.
Using TTC vehicle GPS data as regular traffic probes of road conditions, team TrafficJam Tacos mapped congestion and resulting slow speeds, as well as unreliable areas with highly variable speeds. Based on this insight commuters can make more informed route choices, and planners can target unreliable areas. This baseline opens up numerous possibilities, such as predicting the impact of construction and street events in terms of both geographical reach and duration
Using the Westbound Gardiner Expressway as a proof of concept, the RueView team’s TrafficJam submission forecasts traffic congestion to predict rush-hour periods. RueView can be utilized in various ways, from as simple as avoiding rush-hour to empowering Torontonians to make better transit choices.
The team visualized pedestrian infrastructure and volumes layered together with social media activity to highlight geographical and behavioural hot spots and cold spots. With this, the city has visibility into social/behavioural trends in order to improve, sustain and create new infrastructure. The end-state would include a consumer wayfinding interface that encourages exploring and feedback on municipal efforts to make Toronto more walkable.