With Link headways now around 15 minutes all day, real time data would allow Link’s quality of service to at least match that of a frequent bus route. Unfortunately, that’s not to be.
ST’s David Jackson, answering when we could expect an accurate GTFS feed:
Because of the lead time required for the GTFS data to go live and our very recent headway tweak, GTFS won’t be accurate for Connect 2020 trains.
It doesn’t appear we’ll even have accurate schedules, much less a clue as to a train’s actual position.
As we noted on Twitter, one thing you can do is use the Transit app‘s “Go” function to track your own location on the train. If enough Link riders do this, the app will be able to get something approximating real-time data. Learn more about how Go works here.
The signs at UW Station had no arrival times on Wednesday and Thursday.
Seriously? I would expect them to be able to back calculate how much time it takes once trains leave Pioneer Square. It’s not like there are as many places where delays can occur, especially compared to the southern portion of Link.
The apps “Transit” and “Citymapper” offer crowd sourced times. It seems to work well for the buses, hopefully it can work nicely for Link users.
D’oh! Somehow I missed the mention at the bottom of the article.
I snuck it in there right after publishing. 😉
This has been a consistent problem with OneBusAway. The moment anything unusual happens, the real-time status information immediately disappears, and even the schedule information might not be trusted. Some examples:
– OBA is unable to handle temporary bus reroutes due to construction, even when the detour is known days or weeks in advance (only long-term reroutes that last several months or more, is it able to adjust for).
– When a bus driver goes off-route for whatever reason (e.g. driver misses a freeway exit), OBA completely loses track of where the bus is.
– No OBA data for temporary shuttle buses that replaced Link last weekend.
– No OBA for extra buses deployed for special events.
– OBA shows buses on regular route, even when buses are on snow route. When the emergency snow network is activated, OBA has no clue which routes are running and which ones aren’t.
This is not a complicated problem. Simply having the bus driver run a GPS tracking app with their cellphone would provide robustness against this.
This is not a limitation of OBA but of the underlying CAD/AVL systems that provide the data. Detours are not well handled in any CAD/AVL system. Ultimately it comes down to a decision at what level of inaccuracies do you stop publishing all together. Many administrators choose to only publish real time when there is high confidence in the results and fall back to scheduled times when not.
For the north branch (Pioneer Square to UW Station), there’s a nice low-tech way to estimate the wait time. Just ask the security guard how long its been since the last train left and subtract.
Haha, me asking a security guard that’s funny! Let me know how that works out for you.
No timetable and no real time data makes Link almost worthless. Having to show up at the station 15 minutes before you need to be on a train should not be acceptable to anyone. Instead of spending money on all of the helper people and fancy websites, they should have invested in getting real time data working. It should not fall on riders to correct Sound Transit’s problem.
Riders would have been better served with a shorter complete shut down and shuttle buses which can be run more frequently and tracked in real time.
Extra buses have been added on the most crowded routes serving Link neighborhoods. (How easy is it to track when those extra buses come, BTW?)
There are only so many buses and drivers available, and only so much space to store buses on the bases. Metro could get more drivers by raising wages, at least in theory, but that is a long-term ramp-up requiring months of lead time.
Employers are trying to flex employees’ schedules to spread out the pain, something that has been done, ongoing, but peak-hour travel seems to obey the theory of induced demand.
If the off-peak Link headway could be reliably 15 minutes or less (on both lines), I’d be totally down with a 15-minute clock-face schedule. Heck, a reliable 20-minute headway schedule might still be better than having the train show up seemingly randomly (off-peak only, when there is plenty of capacity).
One additional way to track the trains is an entanglement approximation. If you know how far the north train is from Pioneer Square Station, you have a good idea how far the corresponding south train is from PSS.
It is one thing to run shuttle buses between Sodo and Capitol Hill stations during the weekend shutdowns but it is another to do so between Angle Lake and the UW stations.
First would Metro have that many buses to run that kind of shuttle service and still maintain their regular service and probable buses could not accommodate the number of passengers to cover for Light Rail on the entire route.
It is not the best situation with the 15 minute service and no schedule information available but it is for the short term and you need look at the long range picture that constructions that is taking place now will help provide service to the east side in a few years.
I waited 27 minutes for a northbound train yesterday at 2pm at Westlake, unacceptable. People started giving up and heading back to street level. Expect ridership to suffer.
One note – those aren’t hired helper people, they are ST office staff, all of which have to do a few shifts on the platforms. It’s still money I guess but I think it’s a pretty cool way to keep in touch with rider experience.
I suspect ridership is already down. I live halfway between two stations and my most common trips are UW to Capitol Hill, UW to Westlake, Westlake to Intl Dist, and Westlake to SODO. All those are short enough and have frequent enough bus alternatives that waiting 5-15 minutes for a bus sounds better than waiting 10-30 minutes for Link. Even if I were going to the airport, the 124+A would at least guarantee I wouldn’t miss my flight. And when I say “UW” I really mean the U-District so there’s already overhead in getting to Link, and on the Capitol Hill side there’s also overhead, while the 49 is closer at both ends. So Link is one-half dozen to the other when it’s guaranteed 10 minutes (subject to the usual delays), but when it’s a non-guaranteed 15 minutes it makes me reach for a bus. When Link reaches U-District, Roosevelt, and Northgate it will be a lot bigger advantage for my trips and I’d take it even under these conditions, but for now it’s not. And a lot of people live between the U-District and Little Saigon and have a lot of frequent bus and streetcar alternatives, and likewise between Westlake and Mt Baker. And this inability to stick to the planned 12-minute headways and keep the next-train signs on is probably deterring even more, and will increasingly do so as time wears on.
Yeah, that is what I wrote about before this occurred. One of the key advantages of Link is that it is frequent. Take away frequency (as has happened recently) and bus trips are better — especially at the margins. If you are headed from the UW hospital to Seattle Central, you are going to tough it out, and take Link. But for other trips (e. g. trips from the U-District or to some other part of Capitol Hill), suddenly the 49 looks fantastic. Maybe they should just go back to the old schedules for the buses (run the 43 more often, send the 71/72/73, etc.). I’m kidding, but I’m sure some folks miss those routes this week.
I think Link ridership is getting hammered. I think transit ridership in general is taking a hit. It is the same sort of thing. Trips that seemed really easy a couple weeks ago are now a pain — driving (or calling a cab) seems worth it.
I keep hearing about the old 71, 72, and 73. What areas did they serve?
They were the Expresses from downtown to UW. From there, they diverged. The 71 went up to around 65th and turned East briefly. The 72 went to Lake City Way. The 73 continued on 15th up to Jackson Park IIRC. The loss of the 72 is felt most, as the 372 and 522 don’t cover all of its route. Either the 71 or the 73 still runs, I cannot recall which.
They worked all right, but definitely had their issues. Metro tried to stagger them so you’d get a bus down the Ave every 10-15 minutes, but with different terminals, they’d get bunched, and you’d sometimes have to wait 20+ minutes for a bus. The one mile from 50th St. to campus parkway took as much as 15 minutes (average speed 4 mph).
And, it gets worse. The three branches individually ran every 30-60 minutes, depending on the time of day. When the I-5 express lanes were running the wrong direction, the “express” buses would have to take Eastlake and wait for all the stoplights. And, on evenings, starting around 7 PM, and all day Sundays, Metro would cheap out and not run the 70, instead having the 71, 72, and 73 serve all the stops on Eastlake that the 70 does, making them even slower.
Between downtown and the UW, the 71, 72, and 73 were packed. I had one time at 5 PM on a Sunday where I couldn’t even get on because the bus was too full.
Because of all this, I found myself occasionally taking Uber or Car2Go between the U district and downtown, something that would seem crazy today.
I also used another trick, which still makes a lot of sense today – riding the 512 between the U district and downtown, using the I-5/45th St. freeway station. The 512 finally has real-time info in OneBusAway.
Okay wow, thanks.
We’re talking about ten weeks. Anything volunteers can organize with apps and phones? For funding and authority, really looks to fit category of ‘quakes and terrorism.
Mark Dublin
How about dry-erase boards at each platform, where someone can write down when the last train came, and riders can see the last three or four arrival times before they get erased? The smart-phoned can then go to a website where riders crowdsource those times. Or an ST staffer could consistently upload the info after writing down the time on the board. Kinda like the crowd-sourcing of the color and letter of the day (except those aren’t accepted on ST).
Each station and direction could have its own page, and link to the RTA information for bus alternatives.
ST could also encourage off-peak ridership by only charging during peak.
Exactly what I had in mind, Brent. Don’t believe that “Nothing” is best the system can do- on a lot of levels. Do you know if ST is enlisting volunteers for passenger-assistance?
Agency’s admission it can’t provide train information seems to warrant, if not mandate, that it’s lost a lot of its right to collect fares at all. Factor in the cost of lost time to the system and all its passengers individually, and case can be made that everybody’s already paid.
So also very serious about asking for donations for the duration. The interests responsible for the decade-early loss of DSTT bus operations…did they ever pay anything for the damage to transit?
Might not hurt the Democratic Party to take the lead in initiating and participating. I think that at this point, more than anything else our politics desperately needs opportunities for people to take a hand in making our institutions work.
‘Tis an ill wind.
Mark Dublin
It’s a good idea Mark. If ST could plug phones into each cab and just turn the app on it just might work!
This could have been planned for.
On a related note: why don’t LINK trains have wifi? In this respect Tacoma transit is superior to Seattle’s. All Pierce Transit buses have open wifi for passenger use, and if I remember correctly, it came about through implementing real-time data aboard the buses. Until we get municipal broadband, wifi aboard the trains would greatly help making our time spent in transit productive.
Many (most?) transit systems in the Western World are superior to Seattle’s.
We got a $4bn tunnel we didn’t want instead.
Could say that in a lot of things, Europe’s got a long start on us. In the 1300’s, the Chancellor of Sweden murdered a kinsman on the altar of a Stockholm church now in sight of electric rail same caliber as Link, which in Swedish the Swedes call a streetcar.
And subway entrances too.
Very large difference on the Continent, and the rest of the world, is how recently the average person has been able to afford a car. Whose presence is rapidly making southern Sweden, and also places in Africa, resemble freeway scenery familiar to us.
Wanted or not, the shape of Downtown Seattle, a narrow shoreline flanked by a cliff, fixed it so major ridership generator could only be served underground. But main thing is that to exist at all, only place to work on it is the future. Past is only source of advice.
Mark Dublin
You didn’t want. You can’t speak for all.
@Eric — Seattle did vote against is, so in that sense, “we” didn’t want it.
(“it”, not “is” on my last comment)
Let’s be clear: Seattle transit’s definitely not perfect but we’re doing a heck of a lot better than most parts of the US. For example, my old hometown, Austin TX. What a sad state of affairs — the only rail line (the Red Line, what a name) starts downtown ATX and ends in the distant town of Leander but goes nowhere really useful in between. I remember talking with some Capital Metro officials a few years after the Red Line started running and politely raised a few criticisms, but they got defensive about it rather than being receptive to feedback. CapMetro were quite authoritarian in their decision making at that time and so in 2014 they came out with a plan for more lines that once again, did not address the chief needs of the city. It was voted down handily. Six years have passed, they have not put it to another vote since. Come November they may use the presidential election to finally put a quite limited rail project ($6B) on the ballot which if it even passes will take many more years to be realized. Meanwhile here in Seattle, LINK as it has already existed for some time goes to a score of useful places and more are on their way within a few years. I’m grateful to now live in a city with the political will to build out rail.
But to repeat the question my post posed: why haven’t we got passenger wifi aboard LINK? Would the best way to achieve this be through implementing real-time tracking (as happened in Tacoma)?
I’ve been thinking more broadly about this…could Link handle more people during Connect 2020? I think the answer is yes.
First off, what could be done quickly, given the infrastructure (the center platform at Pioneer Square) and habits (“I probably have to change trains at Pioneer Square) already in place?
My first thought is to close Stadium station entirely to get trains through faster, expecting users there to walk/bus to an adjacent station. (I think it’s the lowest ridership station of the single-track section between Stadium and Westlake.) This might get back to the 12-minute goal on its own, since Sound Transit cites the Pioneer Square – SODO segment as the bottleneck that meant they couldn’t achieve 12-minute intervals.
But can we do better? Perhaps. For example, by closing the Stadium station and going to 15-minute intervals, perhaps a through peak express train could share a track with the Pioneer Square-to-SODO train: UW – Cap Hill – Westlake – Express no stops – SODO – local (likely ending somewhere between Mt Baker and Rainier Beach, but where it ends isn’t key to this plan). Those express trains would perhaps be only 3 cars each, and would likely need to be obnoxiously decorated to make it easy to tell the difference. (and….have yet a new announcer voice or chime/jingle just for that train?) Note that the UW – Pioneer Square train could possibly skip Westlake with this change, but with Westlake’s high ridership and connectivity I’m reluctant to actually recommend that.
But what about rethinking this from scratch? Could we do two through trains every 15 minutes with an aggressive skip-stop pattern in the single-track section? Route A would stop at Westlake – Pioneer Square – Stadium (unless Stadium closed); Route B would stop at University Street – International District. Both southbound trains would go through downtown, then both northbound trains. Some riders traveling within downtown would either walk further or take the bus instead. The downside there is then nights/weekends either has way too much capacity or has a very different service pattern.
If closing Stadium Station during Connect 2020 would have allowed for more frequent trains, they definitely should have done it. The construction work was deliberately scheduled to avoid events at the stadium, and outside of events, there’s no much reason for people to get on/off there.
One potential use for Stadium Station is people with bikes, since bikes aren’t allowed at Pioneer Square and, unlike ID Station, Stadium Station is street level, so no need to wait for an elevator. But, if you have a bike with you, you can just exit at SODO station, instead. At bike speeds, exiting one stop early increases travel time by, at most, a minute or two. There’s a trail running right between the two stations alongside the train tracks.
Certainly the planned logistics of 12 minutes is a failure. It does begin to suggest that reversing a single track first in one direction than another could have worked better. Having a platoon of two trains — one right behind the other — along with a skip-stop operation Downtown (saving time and dispersing loads) could have possibly made a smoother situation.
What I’d really like to see is a panel of light rail operations managers from other places come and give an expert assessment and action plan on ST light rail operations contingencies. They have no one at ST to please (unlike a consultant or internal staff) and could be frank.
Connect 2020 operations are locked in, but closure on a DSTT track will occur again. Is ST ready?
I NEVER thought I would say this, but the streetcar is now the most reliable transit option for my evening commute from Pioneer Square to Capitol Hill. My experiences on Link last week were pretty depressing, and the 12 has good frequency, but in the evening, barely moves up the hill due to I-5 traffic. Walking is, of course, the fastest option now, but not totally desirable in rain/snow, or in the uphill direction.