[simplesleeper]

Transport for London Tech Lead here.

In London, we have been collaborating with google waze - we provide them with information at the local level. Streets that waze should not use for route finding (e.g. roads with schools) can be blocked for use by Waze. Both local councils and TfL can provide this data.

The main difference between Waze and Google Maps is that Waze uses real time traffic data. There are multiple map providers that do this in London. There is a plan to provide different data to different TfL data consumers in the future to try and get map providers to seperate traffic across roads (at the cost of surrendering absolute truth).

London is also host to discussions of taxing roads - this is unsurprising when the roads are mostly paid for by the cost of the London Underground (LU makes the majority of public transport money, but 80% of revenue goes into road, traffic and bus management). The London Assembly (who supposedly keep the London mayor accountable) have written a paper "London Stalling" to suggest road taxing related to mileage across London. This, they no doubt expect, will cut down the traffic. However, from my understanding of embellished models of Braess' paradox, this will only result in similar traffic (on maybe slightly different roads) with an optimum equilibrium between how good a route is and what the cost of it is - meaning people may just take longer to commute and pay more tax.

[Steve44]

> The main difference between Waze and Google Maps is that Waze uses real time traffic data.

Google Maps has real time traffic data and the accuracy is staggering. I occasionally use Google Maps for navigation in the car and it even highlights traffic stationary at traffic lights and temporary road works as I’m approaching.

This level or real time detail really amazed me. I’ve not spent any time looking onto how they do it but assume they are continually tracking everyone with a phone, or at least running Google Maps Navigation, and using their real time location and speed.

[krn]

> Google Maps has real time traffic data and the accuracy is staggering.

But does Google Maps use real time traffic data to calculate the most optimal route in the same way as Waze does? As I understand, that's the main difference between the two.

[eli]

Yes, they definitely both take traffic into account when calculating a route. My suspicion is that they're actually using the same raw data, but Waze is tuned to be more aggressive. Waze routinely guides me off a main road onto a side street to avoid two blocks of congestion for a net gain of maybe 1 minute, Google Maps doesn't.

[fernandopj]

In my experience this is very much true - in Sao Paulo, Waze is heavily used by everyone, and I mostly use Ubers and taxis while working here - and the agressiveness in which Waze will suggest a 5-more-turns, using local roads, to gain (maybe) one minute really sets them apart from Google Maps.

Google Maps' UI is tuned to show you alternatives and their ETA, and it's up to you to take them or not - Waze will just update routes without user input.

[WhiteOwlLion]

Waze aggressiveness is ridiculous. They don't factor in the time waiting in queue to turn left, the intersections you may have to cross causing delays, etc versus just staying the course on a main road. The only benefit are the minute savers that fall for the new route clearing up traffic for me to stay the course. In that way, Waze is great for me while making others take "shortcuts" that really aren't shortcuts.

[Nannooskeeska]

Anecdotal, but there have been several cases where I've been using Google Maps and it pops up a message saying that it found a faster route. All of these cases have been because of accidents or extreme traffic, and all have been on the same route (94 East from Eau Claire, WI to Madison, WI). As far as I know, Google Maps doesn't route traffic through side streets to save seconds like Waze does - it seems to re-route you when there are major delays.

[garblegarble]

>But does Google Maps use real time traffic data to calculate the most optimal route in the same way as Waze does?

Anecdotally in the UK I see Google Maps routing based on live traffic data. I also often get a reassurance notification "you are on the fastest route despite traffic" (and it distinguishes between "traffic" and "usual traffic"). Occasionally I'll also get a recommendation pushed to switch to another route which is faster

[Steve44]

I’ve never used Wayz so can’t compare but quite a few other comments note how aggressive Wayz is at re-routing.

At points where there are multiple route options it will indicate how many minutes faster or slower each option is.

[simplesleeper]

They certainly get different results - but as I said, I don't know the intricacies of their system

[TheRealWatson]

Google Maps does real-time route correction based on traffic delays. I'm not sure about any rules applied for avoiding specific local streets but it definitely warns you about traffic ahead and suggests an alternative route. It also uses historic data to estimate travel times if you tell it when (day, time) you want to leave or when you want to arrive somewhere.

We like to think of all the NIMBY's and school streets but these types of "shortcuts" suggested by mapping apps is also what sometimes leads people into dangerous neighborhoods and tragic consequences.

[jinglebells]

Android location services upload to Google by default. Google even make it a feature https://www.google.com/maps/timeline?pb

[PhasmaFelis]

I've seen Google Maps claim that a specific 500-foot stretch had a traffic jam when there were, in fact, no cars on it or anywhere near it besides mine.

[Steve44]

Just a guess, but perhaps the only other map users had pulled over to the side and stopped. Google may have interpreted that as a traffic jam and you didn’t see them. I’m not arguing a point, it’s just a thought as to how that may happen.

[rahimnathwani]

"when the roads are mostly paid for by the cost of the London Underground"

Taxes on fuel used in motor vehicles (duties and VAT) are something like 30bn GBP/year in the UK.

London has about 1/6th of the UK's population, so let's assume 5bn/year comes from London.

That's 5x the annual operating profit of London Underground[0].

Government income and expenditure is fungible, so it's hard to determine precisely whether X is paid for by Y or by Z.

But it's clear that government revenue from London road users is much higher than the profit from London Underground.

So it seems strange to assert that roads are paid for mostly by London Underground.

[0] https://news.ycombinator.com/item?id=20865111

[zaroth]

Aside from the ~£30b raised in fuel taxes, there's also the ~£6b raised in vehicle excise taxes.

Both of those revenue streams go into a government general fund and are not allocated specifically for road maintenance, because they collect about 3x what is actually spent on highway & road infrastructure each year. [1]

In short, the roads & highways are massively profitable.

London's buses however, lose nearly ~£1b a year, which tanks the entire public transportation budget, and more than offsets the profit generated by the Underground.

[1] - https://www.racfoundation.org/data/road-user-taxation-highwa...

[bmsleight_]

"revenue streams go into a government general fund" Yes but not to TfL, who has to fund this via Tube income.

[zaroth]

TfL actually received ~£3b in subsidies in 2018. That's compared to total gross income from the LU of £2,799m.

> The total of resource and capital grants receivable by TfL in 2018/19 amounted to £3,016m (2017/18 £2,477m). [1]

Total expenditure on roads of £539m, minus Congestion charges of £230m, means that subsidies received by TfL (covered by fuel duty & VED revenues) exceeded TfL road maintenance costs by £2,707m.

The LU does not subsidize the roads. It's actually the other way around. The net-subsidy (a.k.a government's net loss) on the LU is calculated at 9.7p per person-km, compared to a net-taxation (a.k.a government's net income) of 3.8p per person-km for driving.

[1] - http://content.tfl.gov.uk/tfl-annual-report-and-statement-of... (See Page 126)

[edh649]

For reference, from the 2019/20 TFL Budget, direct operating surplus' are:

Streets -£167m

Buses: -£722m

Rail: -£49m

Underground: £823m

Elizabeth Line: -£267m

Other: £9m

Professional Services: -£560m

Property: £69m

Media: £145m

[0] http://content.tfl.gov.uk/tfl-budget-2019-20.pdf

[cucumberfanatic]

"London has about 1/6th of the UK's population, so let's assume 5bn/year comes from London" isn't a good assumption - very few people drive/own cars in London.

[bmsleight_]

The central government does not pass on this income to Transport for London (TfL). So TfL has to fund the maintenance of the major roads (aka Red Routes) in London from the income from London Underground.

[sails]

Slight Tangent: I've got a feeling that private vehicles in Zone 1 (and 2...) benefit a tiny percentage of the population, and should be the next target for reducing inner city congestion (implementation perhaps tricky).

Is this concept something you've come across in your work/field?

If not, any ideas for what is next? I hope to see:

1. Car free days over weekends in high streets.

2. Entire high streets converted to pedestrian, bus & bike only.

3. A move to a more bike/bus friendly city, see Utrecht [1]

[1] Utrecht: Planning for People & Bikes, Not for Cars https://vimeo.com/344373585

[iicc]

https://www.theguardian.com/uk-news/2019/jun/30/kensington-c...

>One of the UK’s wealthiest councils is facing renewed pressure over its decision to block a flagship safer cycling scheme after it emerged the authority could not say what proportion of the people who wrote in expressing views on the plan actually opposed it.

https://www.bbc.co.uk/news/uk-england-london-44405730

>Plans to pedestrianise Oxford Street in central London have been "taken off the table for good", Westminster City Council has confirmed.

[simplesleeper]

1. Has no support from central london residents and

2. Sadiq Khan has already tried to Pedestrianise Oxford Street, but was blocked by parliament.

3. There are cycle lanes and cycle highways and there is a lot of development in the cycle space - and will no doubt be continued. Bus lanes and schemes to improve buses are constantly being tried, but their uptake is low, meaning that it is difficult to increase the service, as they are not very profitable

[genmon]

Small correction: the pedestrianisation of Oxford St (led by Khan and the GLA) was blocked by Westminster Council, not Parliament

https://www.standard.co.uk/news/london/sadiq-khans-plan-to-m...

[neka]

Waze doesn't seem to allow you to specify whether you're a trade vehicle or not so it keeps routing trade vehicles into Regents (and the other Royal) Park which they are prohibited from and liable to be fined.

[simplesleeper]

Agreed - I've had similar issues when renting a big van - I was directed through streets where I didn't comply with width restrictions

[kenned3]

There are GPS tools specifically for trucks, which are aware of width and height restrictions... try one of those next time.

[oh_sigh]

Did you ever tell Google maps that you were in a big fan? Otherwise, how would it know?

[michaelt]

Google Maps doesn't have anywhere to tell it you're driving a large vehicle.

Knowing about and respecting size and weight restrictions is usually a feature on 'professional' sat nav units [1]. Serious delivery companies of course pay the premium to get this feature - but occasional van drivers don't.

[1] https://www.tomtom.com/en_gb/sat-nav/truck-sat-nav/

[2_listerine_pls]

They probably infer your vehicle type from acceleration data

[Piskvorrr]

You could do that, theoretically, from several days of driving, with false positives and false negatives galore. Or you could ask the user...I know, not data driven, boooriiiing, and requires one more interaction.

[gambiting]

How would that be possible? My sister has a new VW Polo with a 1.0L 65bhp engine, the official time to 60 is a glacial 16.5 seconds. I'm sure its acceleration profile would match that of a fully loaded HGV.

[2_listerine_pls]

The time to turn, the time to break, way of turning lanes, all of it could be used to infer a vehicle type. I am not saying they do it, but it's likely.

[seattle_spring]

If this information were used to infer vehicle size, Google would assume everyone in Seattle is driving an 18 wheeler.

[petre]

It's quite error prone. I doubt they do it. Also deriving acceleration data requires sub second speed sampling. This needs to be done on device as it sends too much data sampling over the wire otherwise.

[neka]

Can't imagine that being of any use in London, nothing moves.

[TuringNYC]

Or more easily from the default Bluetooth name

[pfranz]

How far are they willing to go? I wouldn't be surprised London has similar problems, but in LA I owned a house on a "back road" that ran parallel to the 405 going north over a hill. The house was built 100 years ago. Neighbors told me the road was fine until Waze/Google maps about 10 years ago. My small house went up in value quite a bit the few years I was there, but if I wanted to buy a house on a cul-de-sac it would have been double the cost of my house. My neighbors over the 30+ years they lived there had added on to their house (like many others in the area). Last I heard they were unable to sell it--who would buy a large house on a very busy road? (their's wasn't the only example I saw) There's no school or anything to warrant blocking traffic. I'm sure it's a useful road for many people where taking the highway would double their travel time.

They keep trying to expand the 405 and there aren't feasible public transit options. The few that exist rarely put you close enough to where you need to get.

At least in the US, I feel like this road design fell out of favor a decade or two before GPSes and smart-phones. Instead of smaller roads that get improved as traffic increases, they have dedicated roads for thru traffic and neighborhoods become rat's nests, designed to discourage thru traffic. Neither system seems to "grow" very well, but of course locals prefer the "modern" approach.

There was talk in my neighborhood of locals blocking off the road in the middle, which seems to make it more like the "modern" solution I've noticed in medium density housing and solve the "Waze problem."

[cr0sh]

I wonder if they could add speed bumps/tables to slow the traffic down? That might be one solution - and fairly inexpensive, too.

[pfranz]

LA has been adding them all over. I really hate speed bumps. It feels like a band-aid on a design problem that just causes premature wear on your car--but it might be their best option.

[L_226]

> The main difference between Waze and Google Maps is that Waze uses real time traffic data.

Do you mean that Waze uses real time data provided by 3rd parties? Because I have always had the understanding that Gmaps uses location and telemetry data from (the Google maps) mobile apps to determine traffic conditions in near real time.

[roelschroeven]

Waze uses, at the very least, data from people who use Waze. But I suppose Waze also uses the location data Google gets from Android phones. And vice-versa.

[simplesleeper]

I don't know what happens under their hood, but I know for sure they source some data from third parties. Makes sense to, considering TfL provides a lot of data - e.g. live bus times etc

[masklinn]

I expect what they mean is Waze uses real-time traffic data for routing. Google Maps displays traffic conditions but I don't remember that impacting the routing it gives you. Last time I used it, it just marked road sections as high-traffic but didn't truly re-route around on the fly.

[nradov]

Google maps re-routes around traffic on the fly and has done so for many years. It shows you the option and then you have to tap to confirm.

[saltminer]

In my experience it's tap to decline the rerouting, not tap to confirm.

[adrianN]

Do you have a citeable source for the claim that the underground pays for the roads? You often hear drivers that claim more rights because "they pay for the roads".

[buro9]

http://content.tfl.gov.uk/tfl-budget-2019-20.pdf

Page 128.

London Underground does bring in more revenue 2.864bn than it spends 2.041bn, and subsidises Streets and Buses to the tune of 0.889bn (with a little coming from media revenue too).

However it isn't the 80% claimed above.

Time to find out this: about 1 minute.

[soneil]

Drivers don't pay for the roads - any more than smokers, drinkers, or anyone else that opts into a 'sin tax' pays for the roads. Vehicle duty & fuel tax both go into "general revenue", and roads are paid by local councils.

"[road tax] was formally ended in 1936 after Winston Churchill asserted the following:

[I]t is a monstrous assertion that any important body of taxpayers should claim proprietary rights over the particular quota of taxation which they contribute, and that all should not be brought into an area freely justiciable by the House of Commons."

https://www.loc.gov/law/help/infrastructure-funding/englanda...

(aisde; a driver claiming to have more rights to roads because they pay extra tax on fuel, would be like a smoker claiming to have more rights to healthcare because they pay extra tax on tobacco products. I can imagine that wouldn't be popular. "General revenue" is supposed to disconnect source from sink to avoid exactly this.)