Analysis of Public Traffic Control Measures — How the City of Toronto Tries to Cut Traffic Jams on King Street
Problem
King Street, the most heavily used surface transit road in Toronto, moving more than 85,000 people in trams and personal vehicles every day, has become so overcrowded that even the city politicians say that walking on Toronto’s main route is faster than driving. However, even for pedestrians King St. with its busy, narrow sidewalks is not working effectively. Many potential solutions have been attempted to improve the transit service on this street: changes to parking, turning restrictions, increased fines for violations, introduction of supplemental buses to the tram service etc. While these measures have somewhat helped, they haven’t been enough. In the summer of 2017, the City of Toronto started a year-long pilot project to solve the problem of slow moving traffic and traffic jams on King Street.
Proposed Government Solution
During the pilot all traffic on King Street is only permitted to travel one block before being forced to turn right and all of the on-street parking spaces have been removed. The rationale is that one single-occupancy car making a left-hand turn can hold up 100 people on a tram (about the 65,000 people ride trams and only 20,000 people drive personal vehicles). This should also encourage downtown residents to drive less — and around 75% of them are walking, cycling or taking public transit already.
The Implementation Plan
Toronto officials developed an execution strategy comprised of 3 stages:
- Goals and pilot options development (April 2016 — March 2017)
- Evaluation and selection of the preferred pilot (March–June 2017)
- Implementation and monitoring (starts September 2017 and runs for about 12 months)
Stage 1 — Surveying: At the first stage the city committed to engaging stakeholders and the public. To do so the officials organized 4 technical advisory committees, 3 stakeholder advisory groups, 2 public meetings and several surveys and other activities like focus groups, interviews, workshops and public lectures. Public meetings reached 300–500 participants and surveys got feedback from 2,207 people.
Stage 2 — Developing options: the second stage focused on evaluating various options, selecting a preferred pilot and developing metrics for monitoring the pilot program. During this phase, the study team prepared the proposals and sought feedback on the emerging pilot design, and chose the option to limit the driving through the King street to one block only.
Stage 3 — Implementation: the first results showed that the average tram travel time dropped from a maximum of 19 minutes in the morning and 25 minutes in the evening to 16.7 minutes and 22 minutes respectively, a decline of approximately 12%. On the other hand, people started avoiding the King street area. The local merchants are saying that the pilot project is killing businesses because many of their clients prefer to drive to the place they need, but don’t want the hassle of figuring out a way around the city’s new traffic rules. Some media outlets claim that the information reported by the TTC community is incomplete. For example, the reported metrics were measured under the ideal conditions and other circumstances are ignored. There is also some evidence that after the pilot project launch the traffic jams at the nearest streets have increased, the details can be found in the following reports for November, December and January.
The Implementation Budget
While the exact expenditures are not known, the figures are evidently significant given the number of officials involved in various committees and number of labour hours spent on the surveying and preparation of different options. In addition to the direct costs, city also paid for the pilot promotion, more specifically: billboards, radio, website creation and social media marketing.
Consensus AI Approach
The current approach to the problem of traffic jams on King Street involving 2.5 year execution timeline and a large municipal budget may not be the most effective solution. Modern technology allows collection of proposals, surveying and options evaluation to be faster and more efficient and economically rational. We develop Consensus AI with these goals in mind —for the system to be able to take into account the wider range of externalities and factors, achieving the desirable outcomes while minimizing the negative consequences for various stakeholders. Here are some of the things we would do differently:
Stage 1 — Surveying
Consensus AI would be able to help Toronto authorities reduce the time for research and consultations from one year to several weeks and reach a wider sample of relevant city residents via decentralized communities on the network verified through electronic IDs. SEN rewards for providing preference data directly or indirectly would encourage higher public involvement in the decision-making process and improve the AI engine to be able to represent the preferences and interests of each resident in the future based on their previous participation. There is some similarity here to the work of Michal Kosinski, who developed a method to analyze people in detail based on their Facebook activity. For example, Kosinski proved that on the basis of an average of 68 Facebook “Likes” by a user, it was possible to predict things like their affiliation to the Democratic or Republican party with about 85% accuracy.
Improvement in these algorithms could result in even further decrease in surveying times, giving officials near immediate feedback and allowing to understand the opinion of different stakeholders involved — locals versus the outskirts inhabitants, what the pedestrians expect from this change and how it is different from the opinion of car owners, what merchants prefer and what their clients are expecting from the street redesign etc. This is essential in identifying and setting the proper key metrics, since tram travelling times may not be the only measure of success.
Stage 2 — Developing options
Consensus AI engine will automatically take into account all relevant data and simulate potential positive and negative outcomes, searching for optimal decisions based on the available data, other cities’ experience etc., modelling scenarios and comparing different factors. For instance, in this King Street problem the following factors could be analyzed for every scenario: % decrease in average tram travelling time, % drop in number of cars on the King street per day, % increase in number of pedestrians, cost of implementation based on historically similar projects adjusted for current average prices of new road signs etc, number of street businesses to be affected based on how many of their customers are drivers, and % population voting for or against this change.
In addition to driving limitations and parking restrictions, other options could be considered and modelled, for example such as: i) moving the rails from the center of the street to the roadside, ii) making the King street a one-way, like Adelaide St. and Richmond St., iii) prohibiting parking on the King street but providing incentives for building new parking lots in the nearest quarters and allocating any unutilized parking space inside the condominium buildings to non-residential uses to encourage people to continue driving to the city center.
Stage 3 — Implementation
Each proposal could be turned into a task similar to ticket in Jira/ Trello/other task management system. The community and external experts included in the decision process will be able to monitor the task, comment and put forward additional proposals. Communities will also be able to vote and influence, where applicable, the choice of contractors for projects, track the execution and give feedback. In order to provide the transparency of decision making and public funds distribution, the entire process from the idea stage to the final project implementation stage will be recorded on blockchain. This will also allow to reduce the lag between data collection and public reporting. During the implementation stage the AI Advisor will monitor the key metrics and continuously model further improvements.
Analysis of Public Traffic Control Measures — How the City of Toronto Tries to Cut Traffic Jams on… was originally published in Consensus AI on Medium, where people are continuing the conversation by highlighting and responding to this story.
