King County Metro’s RapidRide G Line travels inbound from Madison Valley to downtown Seattle on Madison St. Outbound trips travel in the reverse direction. In May 2025, the G Line had 6,113 average weekday boardings.

The G Line launched in September 2024 and uses the first (and still only) center running bus lanes in King County. The route had a bumpy start as its 6 minute frequencies led to severe bus bunching. Over the past 10 months, these issues have largely been resolved and ridership has steadily increased each month.

G Line Map from Metro

Average Ridership Per Trip

The plots below show the average weekday ridership by stop in each direction, color-coded by time of day. For a more detailed breakdown of how the plots are set up, please refer to the How to Read the Plots section of the article discussing Route 70.

Average Weekday Ridership per G Line Trip: September 2024 to March 2025. “Inbound” is toward downtown Seattle; “Outbound” is toward Madison Valley. Click the plot to view at full-resolution in a new tab.

The G Line ridership patterns show a route that traverses dense neighborhoods with strong ridership all day. Some observations: 

  • Madison St & MLK: The northeastern terminus for the route has consistent inbound boarding counts throughout the day. Outbound departures peak in the afternoon and are slightly lower in the midday and evening. The stop is located in the center of Madison Valley’s commercial area and provides easy transfers to routes 8 and 11.
  • 23rd Ave (inbound)/24th Ave (outbound): This stop pair has the lowest ridership of the entire route. While Route 48 runs on 23rd Ave, there are very few transfers between it and the G Line. This may be due to the inconvenient stop placements that require all transferring passengers to walk at least a block. Additionally, there are not many trips that require this transfer. Someone in First Hill or Capitol Hill going to UW will take Link instead. If they are going to the Central District, they might take routes 2, 3, 4, or 8.
  • 22nd Ave: This stop pair is surrounded by apartment buildings and has consistent inbound boardings all day. Outbound departures are highest in the afternoon and evening. The consistent inbound boardings suggest this line is used for a wide range of trips, such as commuting, running errands, and going out to dinner.
  • 17th Ave: Similar to the 22nd Ave stop, this stop is well used all day. Passengers can transfer at this stop to/from Route 12.
  • 12th Ave: The 12th Ave stop serves the eastern half of the shops and restaurants on Pike St and Pine St in Capitol Hill. Just like the previous stops, it has consistent inbound boardings during the day and a spike in outbound departures in the afternoon. In addition, it has more riders boarding outbound trips in the afternoon and evening. Passengers can transfer at this stop to/from Route 2.
  • Boylston Ave: This is the first stop along the line that has significant inbound and outbound departures. This stop is near the Swedish Medical Center and the western half of the Pike-Pine commercial area in Capitol Hill. The high inbound afternoon boarding and high outbound morning departure counts suggest some employees in the area use the G Line from downtown. Passengers can transfer at this stop to/from Route 60.
  • Terry Ave: The Terry Ave stop is located in the center of First Hill, near the Virginia Mason Medical Center, O’Dea High School, and several hotels. This stop is well used by commuters and students as trips in both directions drop off more passengers in the morning than any other time of day.
  • 8th Ave: The G Line runs as a one way loop on Madison (inbound) and Spring (outbound) between 1st Ave and 9th Ave. The stops at 8th Ave are located near The Polyclinic and several other healthcare offices. The stop’s low ridership can be attributed to half its walkshed being consumed by I-5.
  • 5th Ave (inbound)/4th Ave (outbound): These stops are located next to the Central Library and many office buildings. Unsurprisingly, ridership at this stop follows a commuter-dominant pattern with higher inbound departures in the morning and higher outbound boardings in the afternoon.
  • 3rd Ave: This is the busiest stop on the route in both directions as passengers can transfer here to the many routes that run on 3rd Ave. Additionally, the G Line stops are 1 or 2 blocks from the Symphony Link station.
  • 1st Ave: While it doesn’t quite reach the waterfront, the G Line gets close with its stop on 1st Ave. This stop serves the many offices, shops, restaurants, and apartments in its vicinity. While the stop was likely initially planned as a connection with the now effectively cancelled Cultural Connector Streetcar, it may one day connect with a bus route that runs on 1st Ave.

Daily Totals per Stop

The G Line’s 6,113 average weekday boarding count in May 2025 is lower than many other routes in King County. However, it’s short length and few stops results in impressive per stop ridership.

Average G Line Weekday Boarding and Alighting Counts: September 2024 to March 2025. “Inbound” is toward downtown Seattle, “Outbound” is toward Madison Valley. Click the plot to view at full-resolution in a new tab.

Looking Ahead

As the G Line just opened last year, there are no upcoming plans for the route. Ridership, however, will continue to increase as more of the current residents in the area start riding it. Additionally, the G Line will encourage new development along the Madison corridor, further increasing ridership.

28 Replies to “Ridership Patterns for RapidRide G Line”

  1. I’d like to know the ridership per service hour and ridership per mile of the G Line compared to the E Line and 1 Line, to better contextualize these numbers!

    1. I asked for ridership per service hour too but it’s not available yet. We’ll have to wait for the next annual report.

    2. Adding up the time on the schedule puts the route at 97.7 hours on weekdays. With a ridership of ~6100, that puts the G line at 62.4 riders per hour.

      The E is at roughly 40-45. I don’t have numbers for the 1 line but 1 + 2 is roughly 50-55 (see NTD data)

      1. As of the last service evaluation the A Line had the highest ridership per service hour, with 42.7. Interestingly enough, that was off-peak (just slightly better than on-peak). Given your assessment it is highly likely that the G Line now has the highest ridership per hour of any route.

        It may not set a record though. I haven’t looked through all the old reports, but back in 2015 the D Line had over 80 riders per service hour during peak and over 70 hours midday.

      2. For an estimate on routes that haven’t seen major service changes, you can take the current ridership and divide by weekday service hours in Appendix D.

        For the A line that’s ~9500 rides over 212 service hours, or ~44.8 riders per hour.

  2. I drive the G line 3 days a week. The route is good overall from an operator’s perspective. If I had to make spot improvements, one area that slows the route down in the afternoon is Madison and 5th Ave. A queue jump there would be great so that buses can move over to the middle lane to avoid right turning cars that want to get on 4th Ave. This is only a problem at rush hour.

    1. Thanks for driving. How is the signal on 6th and Spring when headed northbound? A few months ago I noticed that freeway traffic was going before the bus. I don’t know if that’s typical or if that was just the few runs I took

      1. The Urbanist mentioned that section as well. https://www.theurbanist.org/2025/04/16/rapidride-g-line-attracts-major-bus-ridership-gains-along-madison-street/. As I understand it, the problem is this. A bus is heading outbound on Spring (towards First Hill). It passes by Fifth and there is a bus lane to the right. But there are also a lot of cars in that middle lane, trying to turn onto the freeway, in front of the bus. The bus has to wait for its own signal just to go straight. Instead of a two-phase light it is a three-phase light (and the bus can only go straight on one of those phases).

        While a bit of a kludge, I can see how SDOT could give extra signal priority to the bus so that at least it spends less time waiting. Obviously you don’t want to screw up traffic elsewhere but I could see how they could simply allow the bus to go next every time. So if the bus shows while drivers are going onto the freeway, the next phase would be for the bus. If the bus shows up while drivers are going across 6th, the next phase is for the bus.

      2. I wonder if there’s constraints in standard-issue traffic signal software that prevent this. That is, it being hard coded exactly how many signal phases there are, and in what order.

      3. @asdf2 — I doubt it. With signal priority they change the preset phases (making some longer and others shorter). Emergency vehicles are able to completely override them, quickly turning the crossing signals to yellow/red. This wouldn’t be quite that aggressive (obviously) but similar. My guess is you can do anything you want with them. The challenge is in figuring out what you want to do. I’m sure the modeling helps but it isn’t perfect.

      4. Another possibility is that there’s some “best practices” principle that signal phases have to always be in a fixed order, lest drivers get confused and run the red light during a phase they expected their light to be green. Maybe the guideline allows deviation from this for emergency vehicles, so people don’t die waiting for an ambulance to show up, but not for buses.

      5. Again, I doubt it. It is common for a beg button to change the standard cycle and be given priority within a few seconds. Traffic signals have been around a while. A quick search for “traffic signal firmware” led to this: https://fourthdimensiontraffic.com/about/about.html

        The D4 Traffic Signal Controller Software is a powerful package that is best known for its transit priority capabilities and advanced coordinator. The software supports AB3418 and NTCIP protocols and works with a variety of central systems and traffic controllers.

        The product was created in the 1990’s

        Here is the Wiki page for the NTCIP protocol: https://en.wikipedia.org/wiki/National_Transportation_Communications_for_Intelligent_Transportation_System_Protocol. A quick look at the protocol suggests it is quite capable. This ain’t rocket science.

      6. To my knowledge, even actuated signals have static cycle length. What changes from time to time is the length of each phase. The reason that cycle length has to be constant is that in an urban core where signalized intersections are adjacent to each other, signals are usually coordinated with the primary corridor.
        To determine the length of phase, it uses camera or detection loop under pavement to track queue length. If queue is beyond maximum allowable length, it will extend the phase let more cars go in one phase.
        I think the problem of Spring at 6th is that the turns that goes into the freeway constantly hitting its max queue during peak period, so the signal always has to prioritize that phase over other more than ever, even the transit one. This really limits the level of priority G Line can receive. The traffic entering I-5 SB from Spring often backs up to 2nd.
        Spring simply isn’t a good transit corridor. They should have just built the bus lane on the left-hand side of Spring to avoid such turning movement conflict.

      7. In Downtown it’s important to also give time for pedestrians to get across the street every cycle. That’s one good thing about having narrower one-way streets versus an excessively wide two-way street like Mercer. There should never be beg buttons Downtown.

    2. I hope they add that queue jump. I know at least one planner (and me) wanted the buses to run contraflow downtown. This would have made a lot of this a lot simpler. They could still do it but it would be a lot more expensive. They put down a lot of new concrete (making the road harder so that the buses can handle it) and that would have to be done on the other part of the street.

    3. I see the buses from the library, and westbound always seems much slower than eastbound.

  3. Nice work as always with these charts!

    It is cool to see how much more “balanced” a lot of the boardings/departures plots are than some of the other routes that have been featured. Yes, there is a spike during commuting times to/from downtown, but the significant concentration of both origins and destinations near almost every stop contribute to maximizing trip pairs – something we should be focusing more on with our land use & transit planning

  4. Thanks for the piece with weekly G Line ridership.

    Comments.
    MLK Jr. Way East and East Madison Street. This is a common stop transfer. The multiple routes may also impose friction upon one another at the signal and stops.
    22nd-23rd avenues. Some may use the 22nd pair for transfers due to grade. It has a Safeway.
    17th Avenue pair. The common stop transfer also imposes reciprocal friction as the stops are only long enough for one bus at a time. Both Route 12 and the G Line are oriented to downtown Seattle.
    Boylston pair. This may be the transfer for the FHSC.
    1st Avenue. This stop also serves the Marion Street causeway and the WSF Colman Dock. The former routes 10-12 served that connection better. SDOT shifted the G Line to Spring Street from Marion Street to be closer to Link.
    Spring Street. Some are served by the outbound Route 2 trips.

    The various transfers would be more apparent if SDOT and Metro added so-called “spider maps” to the stations.

    Given that the G Line is radial, a more vigorous restructure might have changed radial routes 2, 11, and 12 more than was done and changed more distant and important routes 10 and 49 less. (It is too late to consider the G Line pathway itself; the time for that would have been summer 2018 when the mode choice was made). There may be room to reconsider the nearby routes. There seems plenty of capacity on the G Line to carry more of the load.

    The G Line evaluation should continue and be more intense after network signup data is available. The fall 2024 and spring 2025 network data could be compared with the prior network from fall 2018 (before G Line construction). The post G network could include the G Line and routes 2 (S), 8, 10, 11, 12, and 60. The pre G network could include routes 2 (S), 8, 10,11, 12, and 60. The FHSC and Route 48 were unchanged. It could include ridership and rides per platform hour for each route. The speed and reliability of Route 60 could be checked as well as they elected to keep it on Madison Street after it was reconfigured.

    I expect the Metro operation of the G Line has improved with more operational data. I wonder how many more minutes and buses the G Line takes than forecast by the SDOT consultant.

    1. The friction between at MLK and Madison could be reduced by through-routing the 8 and 11. Right now there are two different routes that share that stop along with the G Line. Thus despite the relatively low frequency on both of them, the combine for about seven trips an hour (four by the 8, three by the 11). The trips are not timed, which means every so often the 8, 11 and G appear at the bus stop at roughly the same time. If the 8 was combined with the 11 and ran every ten minutes it would mean much better frequency for both routes while reducing the conflict with the G.

      The friction between the 12 and the G is caused by poor routing. At a minimum the 12 should branch from the 10 at Thomas. This would provide much better combined headways in this urban area while also avoiding the overlap. There would be a minimal loss of coverage that only a handful of people take advantage of (probably because the 12 alone is so damn infrequent).

      The friction between 2 and G is also caused by poor routing. They share the same general pathway much of the way. From roughly 15th and Madison either bus will do. Heading towards downtown, the 2 manages to stay close to the G the whole way, yet riders never take advantage of this improvement in frequency (to be fair, it is minimal because the G is so much more frequent). Outbound, the 2 shares the same roadway so at least riders can take the first bus that arrives (which is almost always the G). But this also causes the friction you mentioned. Running a bus in this manner is a clear anti-pattern. It doesn’t add coverage. It doesn’t add frequency. It just happens to be going the same way and it shouldn’t. The 2 should be sent up to Pike/Pine and combined with the 10/12 to provide better headways on that corridor.

      With the addition of the 2 on the Pike/Pine corridor it becomes even less valuable to send the 49 down there as well. It should be sent down Broadway to provide much better service on that corridor. Oh, and the 60 should stay on Broadway as well. I would combine routes but a poor man’s spine (49, 60, streetcar) would still be a huge improvement.

      Yes, there should have been a more vigorous restructure surrounding the G Line.

      1. A network is a group of related routes. The G Line was implemented simultaneously with changes in other routes. Together, they are a network. The post G network could be compared with a prior network. What is the change in hours, rides, and rides per hour? An early Urbanist report that the G Line attracted more rides that Route 12 is silly; Route 12 had less service and was running through G Line construction.

      2. What is the change in hours, rides, and rides per hour?

        We don’t have the numbers for riders-per-hour yet. We took a look (as best we could) using initial route data here: https://seattletransitblog.com/2025/02/10/early-assessment-of-rapidride-g-and-restructure/. But that did not look at overall service levels. That gets a lot trickier from the outside. Service on Madison is considerably faster with the G but the G also runs a lot more often. We now have route data but it is tedious to make a detailed analysis. We have considered looking at particular corridors but connecting them together to make a full network analysis is really tough without knowing how the service is spread around. Even when Metro releases the ridership per service hour it is tough. You have various moving targets including a general upward trend to ridership (in the post-pandemic era) along with a shift of service that is not that easy to track. Did RapidRide G take service from the other buses? How much of the increase is due to people feeling more comfortable riding transit? How much of the decrease in service (on other buses) is due to Metro funding and staffing woes (which have nothing to do with the G)? The area is not a self-contained unit and even if it was it gets tricky.

        In my opinion the Rapid Ride G Line is a success. It cost a fair amount of money, but my guess is a lot of that money was spent on things the city would have needed to spend money on anyway (like fixing sewer lines). The buses run quite often, and yet the ridership per service hour appears to be outstanding. Despite the various flaws (they should have run contraflow downtown) and various hiccups, it is still pretty damn fast for a bus route that runs in the middle of the city. Various naysayers who talked sh** about it when it first opened are remarkably silent now. This is in contrast to those of us (myself included) who continue to talk sh** about the streetcars because they are clearly flawed (each in its own way). The G definitely isn’t perfect but it is a good example of what you can do when you focus on right-of-way, frequency and a straightforward route. It would be even better with a network that did it justice.

        Which makes the numbers even more impressive. These numbers are very good. But they aren’t nearly as good as they would be if buses like the 2 were moved and other buses played more of a complementary role. While I like to talk about the G poaching riders from the 2 the opposite has happened as well (although not as much, obviously). If I’m just trying to get up the hill and the 2 arrives, I’m taking it. Sure, the G will be there arrive within five minutes — who cares? The 2 is here *now*.

        Thus ridership on the RapidRide G looks a little like Link ridership would have looked like if Metro decided to ignore it. Very good, but not nearly as high. Not only that, but the decision to largely ignore it has resulted in a network that isn’t nearly as useful and has less overall ridership as a result.

  5. This stop pair has the lowest ridership of the entire route. While Route 48 runs on 23rd Ave, there are very few transfers between it and the G Line.

    It is all relative. 150 boardings (inbound) at this stop isn’t much for the G Line, but it is quite a bit for a typical bus. Now consider the transfers. If you are coming from the south it is a hairpin turn towards First Hill (you would have been better off taking the 3/4 when you had the chance). It makes the most sense coming from the north. But if you started at the UW it is probably faster to take Link as the first step (as you mentioned). So it really only makes sense from Montlake towards First Hill. This is a fast way to get downtown (or get to First Hill).

    So why don’t more people take it? Partly it is the transfer, but mostly it is because there just aren’t that many riders in Montlake. The 48 is not very frequent. Montlake is relatively low density. This leads to fairly low ridership on the 48 for this stretch. Between UW Station and John the bus picks up 268 riders. Not bad, but not great either. About 174 get off the bus at John. There are only two stations with more alightings (on a southbound 48): 23rd & Jackson and Mount Baker Station. Thus this is stop has high ridership from the perspective of the 48.

    There are three likely transfers from here:

    1) Take the 11 towards Madison Park. The 11 is even more infrequent than the 48 and there are only so many people headed that way. My guess is only a handful do this.

    2) Take the 8 to the west.

    3) Take the G Line towards First Hill and downtown.

    While I’m sure there are plenty of people headed towards the Thomas/John/Denny corridor, I think it is safe to say that more people are headed towards First Hill and Downtown. Thus it is quite likely that well over a hundred people transfer here. That still isn’t that many — running the 48 more often and/or improving the transfer would help.

    1. I believe there may be some evidence of the G being used for Madison-UDist trips. Both the 43 and 48 (but critically not the 8) have seen large ridership increases (37% and 24% YoY respectively) starting a few months ago. The only explanation I can find right now is that people have started to realize using the G is easier than taking the 60 or 49 or their feet to Link.

      1. I would also say that people are likely transferring on the 22nd Ave stop, not 23rd because otherwise youre backtracking which adds time

      2. Both good points. It is not always easy to get to Link. As for the transfer, it seems a bit tricky. Northbound, riders probably use 22nd. It isn’t much closer if you stay on the bus longer.

        In contrast, I think a lot of southbound riders would use the bus stop on 23rd. It is a pretty short walk: https://maps.app.goo.gl/PX6HzBYS9Q4EuRx57. You are backtracking, but not very far. Then again, I can see why someone would walk the other direction. The problem is the traffic light. It often favors other traffic. So basically you get off the bus, walk to the corner and wait. Eventually the bus shows up and picks up the riders waiting there. It moves a few feet closer to the intersection. The light turns green and you are allowed to cross the street just as the bus crosses going the other direction. You can wave and holler all you want — the bus isn’t stopping.

        Walking to 22nd is farther, but you only cross one intersection with a traffic light, and it looks like a pretty easy one to jaywalk. Even if you don’t jaywalk, it is likely you could probably catch the bus by walking quickly or jogging. So basically you get off the bus and start walking towards downtown. All the while you look over your shoulder to see if the bus is approaching. It is, so you walk faster. You reach Denny Way at roughly the same time and the light is red. You jaywalk or wait for the walk signal and then jog onto the bus.

        I could also see riders doing a combination. You get off the 48 and look for the walk signal (and bus) as you reach the intersection. If you can cross 23rd (and the bus hasn’t arrived yet) that is where you head. If not you walk towards 22nd.

        A lot of the issues surrounding the G Line were tricky but I think they simply made a mistake here. The bus stop should be on the westbound side of 23rd, where there is a bus lane: https://maps.app.goo.gl/vMXGaXQjahJuLgQV6. That would make this transfer really easy. Just get off the bus and round the corner. The stop at 22nd would be a bit close but it could always be moved a block closer to downtown (between 19th and 20th). Going the other way is trickier.

Comments are closed.