This is a repost from 2013 in the issues with the TomTom congestion report of which the latest version has been released today.
TomTom have once again released their meaningless congestion index.
TomTom has announced the results of the TomTom Traffic Index 2013, revealing New Zealanders waste up to 93 hours a year stuck in traffic and that Wellingtonian’s experienced the worst traffic delays during peak hours, spending up to an extra 41 minutes in an hour commute. The Index also revealed that traffic congestion on non-highways is worse than main roads.
The regional results of the Index covers 9 major cities across Australia & New Zealand, with Sydney listed as the most congested city in the region, followed by Auckland and Wellington.
- Sydney 34%
- Auckland 29%
- Wellington 28%
- Melbourne 27%
- Perth 27%
- Christchurch 26%
- Adelaide 25%
- Brisbane 23%
- Canberra 17%
According to the TomTom Traffic Index, Friday morning is the least congested time to commute in New Zealand. The most congested commute was found to be Tuesday morning, and Thursday evening.
There were no cities from our region featured in the top 10 most congested global cities. Auckland, Wellington and Christchurch ranked 22nd, 25th and 42nd respectively in the world’s most congested cities list.
The ranking by overall congestion level in 2013 were:
- Moscow 74%
- Istanbul 62%
- Rio de Janeiro 55%
- Mexico City 54%
- São Paulo 46%
- Palermo 39%
- Warsaw 39%
- Rome 37%
- Los Angeles 36%
- Dublin 35%
“The TomTom Traffic Index gives us a great insight into the state of our traffic network. By providing an accurate analysis of traffic flow and guiding traffic away from congested areas, TomTom plays a key role in helping to ease congestion, improving the traffic flow for the cities,” said Phil Allen, TomTom Maps and Traffic Licensing, SE Asia and Oceania.
It’s meaningless for a number of reasons including:
1. It measures the difference in speed between free flow and congested periods. That means cities with lots of all day congestion there isn’t as much of a difference between peak and off peak times and therefore they get recorded as having less congestion.
2. It doesn’t take into account the speeds at which roads most efficiently move traffic – which is not in free flow conditions. This is something picked up on in research conducted for the NZTA by Ian Wallis and Associates
Various definitions of congestion were reviewed and it was found that the concept of congestion is surprisingly ill-defined. A definition commonly used by economists treats all interactions between vehicles as congestion, while a common engineering definition is based on levels of service and recognises congestion only when the road is operating near or in excess of capacity. A definition of congestion based on the road capacity (ie the maximum sustainable flow) was adopted. The costs of congestion on this basis are derived from the difference between the observed travel times and estimated travel times when the road is operating at capacity.
The graph below shows the engineering definition mentioned above.
3. It doesn’t represent all trips on the transport network. We know that even though only about 10% of all trips to work (which excludes trips for education) are made via PT, it still represents a lot of people. For trips to the City Centre more than half of the people arrive by means other than a private vehicle and many of the PT users arrive via the train, ferry or a bus that has travelled along bus lanes. The people on those services or walking/cycling are doing so often completely free of congestion and so their experience isn’t counted.
4. The data only comes from people with a TomTom device and who have obviously had it on. Many people making the same trip on a daily basis or running a regular errand like going to a supermarket are likely to simply leave their GPS systems off. That is likely to distort the overall figures as they may use routes that have less congestion on them than the route the GPS would select.
5. It can disproportionately impact on smaller cities. As an example if you’re in a larger city and have a 45 minute commute however congestion delays you by 30 minutes that equates to a 67% congestion rate however if you are in a smaller city and you’re commute is only 15 minutes and you get delayed by 15 minutes that’s a 100% delay despite the hold up being half of what the bigger city experienced.
It’s starting to get a bit old now however there’s a good piece on the issues with the methodology in this piece from Reuters, some of which is covered above.
Lastly in the email I received about it they also mentioned this
Of the 138 countries surveyed for the Traffic Index, a global average congestion rate of 26% was recorded, placing New Zealand above the average with a rate of 28%. To put things in perspective, Wellington and Auckland even beat out New York City (39th) in the global rankings, a thriving metropolis of 8.4 million.
So we have worse congestion than New York, a city where the majority get around by methods other than a car and who in recent years has been reclaiming road space for pedestrians, cyclists and buses. Perhaps we should do more of that.
Lastly if we really want to move people around then then the Congestion Free Network would allow people to do that completely free of congestion giving some real choice.
Hopefully i get to see a proper working congestion free network in my lifetime.
That 2030 is very unlikely. With the speed Auckland is at, its more like 2060-ish. Lets just hope CRL finishes in time.
wow no comments yet, shows how meaningless their data is….my question is does anyone know (without me googling it) how this compares to Googles “traffic” colour overlay on their maps. I’m guessing it is just absolute speed, and does it gather data from apple devices or just Androids?
It’s all about perception. The road’s users (i.e. real people in the real world) define congestion as anything that increases their personal journey time. Engineers consider congestion to be when the road has reached saturation (i.e. reached maximum flow). The road might move 150% more vehicles if they travel 35% slower (i.e. at 100% efficiency), but all the road users see is a 35% increase in their personal journey time. The same people who complain when their internet goes slow when everyone streams GoT or “The Batchelor”. 😉 Not everyone is prepared to make that personal travel time sacrifice for “the common good”…
Listening between the lines to the fans of driverless cars I get the impression that they will all miraculously operate at the extreme right, the most optimal part, of the speed vs flow curve.
Using Waze data ( mobile app) it is easy to see where the congestion is and what time of the day.
I say, cycling is the answer. Better and more seperate cycling paths
Its alarming its reported as news, its really just sponsored content and the article should contain a warning as such.
Does anyone use tomtom’s products in the age of smartphones, google maps & Waze etc?
Seriously there sample size must be getting pretty small.
According to https://en.wikipedia.org/wiki/TomTom TomTom have their own mobile app, supply maps and traffic data to Uber and Apple, and partner with Mazda, Renault, and FIAT Group.
I get the criticism/limitation of the TomTom data, however as to your point (5), isn’t that a feature rather than a bug?
If your small time NZ town commute goes from 15 minutes to 30 minutes due to congestion, any way you look it it, that is a damn site worse than a commute in say NYC going from 1hr to 1hr15mins.
The fact NZ towns under any metric can rank 22nd, 25th and 42nd respectively in the world is a disgrace. Last time I checked these cities were nowhere near 22nd, 25th and 42nd in the population stakes.
As others have pointed our the methodology is very much from a user experience perspective rather than looking at any engineering traffic flow dynamics.
Would there be any gain from reducing the speed limit during peak hours?
Would a reduced speed limit also reduce the TomTom perceived congestion rate?
On behalf of TomTom, I would like to respond to your comments.
TomTom has over 500 million devices which is increasing significantly year on year giving us travel speed every day and typically these devices are updating every 1 – 30 seconds. The devices include not only TomTom GPS, but also from third party Mobile devices, Mobile Apps, Integrated in Car navigation systems, taxis and Fleet Management solutions. A mobile device does not necessarily need to be running an app to provide the data if TomTom has an agreement with that manufacturer however rest assured that all data supplied is totally anonymous and cannot be linked to a user. With this very high volume of data we can measure traffic speeds and therefore congestion very accurately. In Auckland’s case, we monitored 7,613 km of directional roads over 365 days and collected 16.7 million kilometers of driving data. This very high volume of data provides the credibility to the results.
Now in response to your points above, I have had to do these in individual blogs:
1) Measuring between Freeflow (typically 3am in the morning) and live traffic flow is a well recognized means of demonstrating the extra travel time incurred to a driver and that is what we are making people aware of rather than trying to define “congestion”. I used Bangkok above as it was rated our 2nd worst city out of 390 and was by far the highest city for evening peak traffic and yet its freeflow is really freeflow at 3am in the morning allowing us to provide an accurate determination of the additional time a driver will incur in driving at peak times (again not using the word congestion). We could have used the Legal Speed Limit as an alternative measure but we believe Freeflow is more appropriate approach.
2) TomTom is measuring the speed of vehicles on “All roads” 24×7 for 365 days. Therefore with the volume of data that we are getting (see above) we can derive very accurate travel times on every road segment across each day of the year and so we can derive the delay on any particular journey for a particular time. This then allows us to derive the “average” extra travel time that you can expect to incur on a journey in that city. This is what the traffic index is. So in Auckland you can expect to spend 38% longer in 2016 in a journey on “average” compared to travelling in freeflow.
3 & 4) Because TomTom sources data not only from its own devices but from many 3rd party sources as well including mobile devices, mobile apps, portable navigation systems, in-car navigation systems taxis fleets and Fleet management systems, this provides a very robust cross-section of vehicles on our roads. A majority of these systems are on “all day” and this can be seen in the volume of data coming in every 30 seconds on most roads across NZ.
5) TomTom is measuring the additional travel time on your journey across every road segment irrespective of the length of the journey or the size of the town.
I hope this helps to better explain the traffic index results and corrects some of the wrong assumptions that have been made above
Great to hear from you to clear up those misunderstandings. I have a few questions with regard to your methodology.
Do you compare free flow to peak hour, or peak periods, if so which periods?
If you are selecting peak periods, why? Your system doesn’t seem to account for the all day congestion that occurs in many European Cities.
Why did you choose to primarily express congestion as a ratio rather than an absolute value, surely a city in which residents face 10 minutes of delay on a ten minute trip should be considered less congested than a city in which residents face a 15 minute delay on a half hour journey as most motorists are more concerned with the amount of delay on a commute not the share of it?
Do you compare data from TomTom to actual traffic counts and scale your data accordingly where it is a higher or lower share of vehicles than average?
How do you account for very small sample sizes for freeflow periods in your data when they occur?
Have you considered producing a report for each city to show the congestion across cities?
“So in Auckland you can expect to spend 38% longer in 2016 in a journey on “average” compared to travelling in freeflow.”
Sure, but the critical failure is you standardise to an arbitrary minutes per hour based on a percentage difference, without even stating what the average time actually is at peak, or at freeflow.