This is a guest post by Pshem Kowalczyk, a long-time follower of the blog.
With great fanfare, just over six months ago (on 12 November 2023), AT launched its interim busway for the NorthWest region, with the new WX express service at the heart of the changes. I live in Te Atatū Peninsula, so these changes have impacted me directly. I realised I tend to catch the bus much more often these days, although I still predominantly commute on my e-bike.
The NW interim busway solution comprises a number of key routes:
- WX1 – between city centre and NorthWest shopping centre, travelling along the motorway between Newton Rd and Hobsonville Rd
- 13 – between Henderson train station and Te Atatu Peninsula
- 11T/W – between city centre and NorthWet shopping centre, travelling along Great North Road, the motorway between Pt Chev and Lincoln Road and Waiumu Rd (11W) and Triangle Road (11T)
A number of other routes that intersect with the busway services were also either updated or created.
The “busway” as such is made up of a number of stretches of bus shoulders (mostly between Pt Chevalier and Westgate), but it lacks continuity across the motorway interchanges, and critically – there’s almost no bus priority between Pt Chevalier and Newton Rd.
[Editor’s note: over the last decade and a bit, despite massive investment spent widening SH16, NZTA egregiously and repeatedly threw NW public transport under the bus instead of building a proper facility akin to the Northern Busway, This ranks right up there as one of the most regrettable planning failures for Auckland.]
Auckland Transport provides a GTFS (General Transit Feed Specification) API, which allows for querying things like real-time departures or bus occupancy. So, I decided to tap into that feed and started collecting information about bus movements in order to figure out how well the interim busway system is performing.
It’s important to acknowledge that the analysis that I’ve done has some limitations:
- Some bus runs are not being tracked at all, hence their performance is not reflected here
- Occasionally I saw “glitches” – buses disappearing from the system (perhaps due to a bus failure or other incident)
- AT’s system doesn’t store “historic” data, so once the bus run is finished the data disappears, if my system didn’t manage to query the data in time – it also would be lost.
And a note: you can click on any of the graphs below to open them at full size in a new tab.
Journey times: how long does it take?
The first thing I looked at is how long it takes for the buses to travel between various bus stops. I used the data from February and May (workdays only). The graph below shows the WX1 travel times between NorthWest (Stop 5959, Kedgley Road/Maki Street) and Karangahape Road (Stop 7135, Rainbow Bridge/Karangahape Rd).
The main takeaway here is that the actual travel times vary a lot throughout the day as a result of congestion and are on average 10-12 minutes shorter than the timetable suggests, even at peak. Also between February and May, travel times improved a bit.
In the westbound direction, between Karangahape Road (Stop 7136, Rainbow Bridge/Karangahape Rd) and NorthWest (stop 5941, NorthWest Shopping Centre) the WX1 travel times look like this:You can see clearly here that westbound services get caught up in traffic congestion: it takes 50% longer for the run during the afternoon peak than during inter-peak. In this case as well, the timetable is more pessimistic than the actual run times. The travel time in that direction is getting worse over time, too, as we can see in the graph below which shows median travel time by week (during the afternoon peak):
For comparison, here’s the travel time for the NX1 service between Albany and Victoria Park (the data covers the same days as for WX1 above):
Even though the outbound travel times are not exactly flat, the variation is fairly small – the NX1 only takes about 4 minutes longer at peak than at other times. What’s interesting here is that the timetable is definitely more optimistic than the actual travel times.
Reliability: how close to the advertised timetable?
As a user of the bus system, I also want to know how likely it is that my bus will be at the bus stop at the advertised time. So I looked at a number of bus stops and routes to see how well they actually perform. The key routes run with the following nominal frequencies:
- WX1 – every 10 minutes, peak and interpeak
- 13 – every 7-8 minutes at peak, 10 minutes interpeak
In theory, these frequencies should allow for a “turn up and go” type of service.
First, let’s have a look at the most frequent route – the 13. I checked the time between buses for two of the stops: 5102 – Roberts Road in Te Atatū South and 5080 – Abbotleigh Avenue Te Atatū Peninsula (data collected for workdays in May):
These two graphs show the distribution of the amount of time between buses, i.e. the waiting time for a person at the bus stop. A few things can be seen here:
- In about 75% cases (the big rectangle in the middle of each column plus the bottom “whisker”) the wait time is quite close to or even less than the advertised timetable
- The reliability visibly drops at the later stop (Abbotleigh Ave)
- The buses tend to bunch up at peak time (note how the bottom “whisker” reaches 0 minutes between buses)
- There can be about 20 minutes between buses in the worst case scenario
But all of that means that overall the service is actually quite good.
Let’s now look at the flagship service: the WX1, at Te Atatū Interchange stop A (towards the city centre) and at the Rainbow Bridge on Karangahape Road (towards NorthWest):
The graphs above show that 75% of buses show up around the advertised time or earlier, with half of them arriving within a few minutes of the schedule. The afternoon (17:00, 18:00) buses from Te Atatū Interchange towards the city centre seem to be the most unreliable, perhaps as a result of the outbound services being delayed. There’s some bunching up happening throughout the day as well.
But again, this service actually looks quite good.
Here, I couldn’t resist comparing the frequency and reliability of the WX1 at Te Atatū with the NX1 at Akoranga.
This comparison does show that the WX1 has some way to go before it achieves the frequency and arrival time reliability of the NX1. The graph shows that in the worst case scenario the wait time for an NX1 bus is never much longer than 10mins, whilst for the WX1 it’s easily over 20 mins.
Occupancy: how busy are the buses?
The next thing I decided to check was bus occupancy: how full are the western express buses? AT reports occupancy as a single number, from 0 (the bus is empty) to 5 (the bus is too full to pick up passengers).
I checked what occupancy looks like for the WX1 at Te Atatu Interchange stop A (citybound) and at the Rainbow Bridge on Karangahape Road (Northwest-bound).
If these numbers are to be trusted, it seems the likelihood of being left behind, even at peak, is not huge. I’d be interested to hear of people’s experiences with this.
Over the period the WX1 has already been operating, the mean occupancy looks quite stable. This graph shows morning peak for the citybound service at Te Atatū Interchange:
Where are the congestion pinch-points?
Finally, I checked where the biggest pinch points are, by comparing travel times between stops during peak time (7am-9am or 4pm-6pm) with the inter-peak times (10am-2pm).
The graphs below show how much longer the travel is likely to be at peak, with the especially bad (or unreliable) stretches highlighted. What I found quite striking here is how long can it take for the WX1 to simply get from Lower Albert Street to Victoria Street West in the afternoon peak.
AT has recently made some improvements to Te Atatū Rd, and from mid-May the T2 lane in Te Atatū South became operational, making it easier for the Route 13 buses to travel along Te Atatū Road. The graph below shows the travel time between Henderson Station and Te Atatū Interchange in May, before and after the T2 lane became operational.
As you can see, the new bus lane saves about 3 minutes on each bus run.
Another thing I wanted to verify was the claim at the Te Atatū Interchange that the citybound 11T/W takes about 5 minutes longer in peak than the WX1.
Which turns out to be true: and off-peak, the difference is closer to about 8 minutes.
Finally, I decided to settle a long outstanding question – which day of the week is the worst when it comes to travelling from the NorthWest to the city. This graph is based on all the workday data from the beginning of the year, and shows the morning peak travel time from the Northwest to the city.
It’s clear that for travel time, both Wednesday and its close runner-up Thursday are not great (or conversely great days to take the e-bike instead, if you can):
As for the return journey in the evening peak, city-to-NorthWest: Thursday just edges out Tuesday and Wednesday as the longest bus ride home of the week.
Interestingly, Friday comes out best in both directions, with Monday second-best. What’s that about?
To summarise my analysis – and to answer the question in the title – quite good, actually! The interim busway seems to be performing reasonably well, despite the fact that buses regularly get caught up in traffic. The high frequency and relatively good reliability make this service a great addition to the network. Let’s hope more priority measures can make the services even better.
Great data work! My main takeaway is that WX1 and NX1 are almost identical distances, and WX1 takes twice as long in peak times.
Confirmation that West Auckland, while taking the brunt of intensification (now and planned: https://www.aucklandcouncil.govt.nz/plans-projects-policies-reports-bylaws/Documents/future-development-strategy-summary.pdf), gets the shaft on public transport infrastructure.
Good point.
Average speed should be an important metric for bus service performance. Speed is capacity by stealth.
I actually did look at the average speed as well (end-to-end runs),
WX1 citybound (morning peak): 19.98km/h
WX1 citybound (average): 23.8km/h
NX1 citybound (morning peak): 37.8 km/h
NX1 citybound (average): 39.99 km/h
Which gives you a very good idea on the performance differences.
Incredible stuff. It’s no wonder we can beat the bus on e-bikes. 20km/h
Nice, so roughly it’s about half as good as the NX buses. Looks like major speed increases can be had by resolving the pinch points.
During winter I pedal a commuter bike, with heavy frame, mudguards, fat squishy tyres, etc. I am overweight and old, but still average faster than 20km/h and the peak hour bus.
I was riding on the cycleway past the buses on the motorway last night thinking about how they were about to be completely held up around St Lukes and further west, and wondered if the passengers pay any attention to the bikes going by.
Lots of reasons cycling is better than the bus.
Just had a look at the 25B bus from Mt Roskill to Civic Centre, this is scheduled to take 32 mins to do 7km = 13km/hr in morning peak. So the North West actually has a much better service than the Central Isthmus despite popular belief.
For comparison, I did just look at the 25B (end-to-end) from Blockhouse Bay to Civic (over last 4 months):
– average across the day – 22.49 km/h
– morning peak (7-9am) – 20.14 km/h
what’s important to note that during periods like school holidays the morning peak average speed can climb all the way up to 28 km/h.
It’s very true. The quality of that interim solution is laid bare here. It works, but at peak the performance tanks badly, and what’s worse – its clearly getting worse (since the buses get stuck in traffic)
And it impacts people really badly on the Westbound journeys – even in the AT modelling they recognised people could easily be waiting at the Lincoln Rd interchange for 30 mins+ late at night.
And of interest riding through the TAT interchange I noticed a number of AT staff ( the sort that wear suits) on some kind of visit, spotted the Project Lead for NW Rapid Transit amongst them
But as stated above, the interim solution is still much faster than what is available on the Isthmus. We need to start with these cheap interim solutions everywhere instead of gold plated solutions to some lucky spots.
It’s a nice journey to Karangahape Road but a hellishly long trip down K road and Queen Street. Plenty of scope here for AT to improve things. And coming back up Albert Street, Wyndham and Hobson takes far too long and again, there are plenty of opportunities for quick wins here.
K Road would be an easy fix if AT would commit to a 24/7 bus lane. Not sure what the fix is on Queen St though. It’s already pretty bus friendly, short of totally banning cars (which is far too controversial to get over the line).
Probably wouldn’t be too hard. They’ve already got a few sections bus only, so adding a few more areas (e.g. Victoria to Wellesley when Wellesley becomes bus only as there is nowhere a car could go/come from) wouldn’t be that hard.
Also, Queen St from K Rd to Mayoral could have a bus lane added easily. Is already 2 lanes each way.
A big problem with Queen is the amount of lights that have pedestrian-favouring timings, which is not something I think should be changed.
Anyway, is Queen St not temporary? I thought the plan was to run the NW buses down Albert St once it’s fully reinstated
Huge issues with Wyndham Street and then K Road. I expect a lot of the delays will ease up as the City Rail Link work finishes and buses won’t be using Wyndham Street to get to K Road. Auckland Transport also needs to harden up and get on top of cars using K Road – it is either a key public transport corridor or a key arterial route for private cars. It can’t be both. It is absurd that it now takes 45 minutes to get into the city from suburbs like Avondale / Pt Chev because there are no bus lanes on K Road aside from during peak hours. It should be all day every day.
Great work Pshem, one aspect that my daughter brought to light for me was the new alternatives the changes have offered. She and her friends can go from Te Atatū Peninsula to New Lynn or Westgate easily- in the past both destinations were via Pt Chev (!) or Henderson.
Nice analysis. Auckland Transport or Waka Kotahi should hire you.
Mondays and especially Fridays have always had the least traffic because more people take those days off for long weekends, even before Covid. Now that people are WFH more that trend is even stronger. When I was a traffic engineer, we used to collect traffic counts only on Tue-Thu.
Yes, or they leave work early on Friday etc.
The majority of people in my office (Wynyard Quarter) work from home on Mondays and Fridays.
The answer to your “what’s that about?’ question is surely simple?: working from home
Not everyone is a computer worker, and not every computer worker is offered the opportunity. And not every computer worker offered the opportunity should take it up.
WFH helps and is sensible (I do it myself several days a week) but it doesn’t take away the need for a good (public) transport network…
Noticeably more workers WFH on Monday and/or Friday off than other week days. I’ve just changed my WFH day from Friday as I’ve stopped been an interim manager (being easily available to team members is useful, Friday has fewest in the office). I’ve found the lighter traffic gives me a very good chance to get home in time for an unhurried dog walk before going out.
It only takes a reduction in 10 % of traffic to have a significant reduction in travel times.
I don’t think anyone is suggesting we don’t need PT because some people work from home.
This is great analysis and something I have been wondering myself catching the bus from Te Atatu South to Karangahape. Two (anecdotal) observations: first, your data bears out my suspicion that the WX1 is faster than the 11 in-bound, but I believe the 11 may be slightly faster outbound than the WX1 – would be great to see this data too! And 2. The timing of the 13 and the WX1 doesn’t quite align – I seem to constantly just be missing a WX1 and wonder whether the transfer would be smoother if AT tweaked the 13 timetable so they’re leaving Henderson a couple of minutes earlier.
Ebike to work takes me 40 min door-to-door so remains top-tier commute as long as it’s not pouring with rain.
Looking at the data for the outbound 11T/W vs WX1 I can see a very similar pattern to the inbound one. WX1 is always faster, might be only by about 3-4 minutes at its worst time (4.30pm), but faster none the less. At other times the difference is around 5-6minutes in favour of the WX1. Throughout the day (till 2pm) the real future of actual dedicated corridor shows – 12 minutes of travel time between K’rd and Te Atatū Interchange.
Also worth flagging that beyond Westgate, the future gets even murkier – the Brigham Creek roundabout is, from what I understand, a totally separate project to anything happening along the North Western up to Westgate.
By the time anything happens there, there will be infill housing all the way from Westgate to Huapai and it will effectively be one long continuous urban motorway. That’s despite it already being the main interchange between SH18 and SH16 and the traffic coming through Whenuapai already triggering the need for a road upgrade which is currently being built. But the question of the Brigham Creek roundabout itself remains.
There’s little chance of rail returning (frustratingly) to Huapai, so the clarity around the future of PT along the corridor to Brigham Creek is massively urgent and already several years overdue.
Do these buses have all door boarding? Surely that would speed things up particularly on the double decker buses on NX1. Although the tagging off is super painful
I definitely miss the London approach where all you had to do was tag on and that was it – 1.50 pound fare if i recall
Agree, imagine the simplification if it was a flat fare. Also the increase in ridership from the further suburbs that are currently quite expensive.
The pinch ‘points’ – appears massively let down by the very limited shoulder lane priority for buses. NW Rapid Transit is still far too far away. Travel time saving from Brigham Creek Rbt to at least Rainbow Bridge is critical to mode shift, as any reduction on car travel time needs to be outweighed by RTN saving, to maximise mode shift. Lots of cars need to be dumped in a paddock near BCR Rbt for the day.
Given that the replacement buses on the new WX1 contract are to be electric double deckers,there needs to be a lot of work done to improve loadings. On these figures,they,the double deckers would be at about one third capacity. New computers run slow on copper broadband.
Great data work. Are there any other little upgrades in the pipeline like the T2 lane in Te Atatū South?
There were some talks of giving the WX1 priority on the on-ramps at Newton and Hobsonville Rd, but it’s been months since I heard anything.
Waka Kotahi finished widening of the Te Atatū city-bound onramp to four lanes (bus lane, T2 and two traffic ones) a few weeks ago, that helped a bit at peak as well.
I know that there was discussion of fixing the merge point at Royal road (city bound) to be similar to the one at Rosebank, so the buses don’t have to merge into general lane before the onramp, but apparently that requires some infra changes (stormwater?) and I’m not sure if that’s even planned.
Anything else would require major changes.
OK thanks. I had mixed up that city-bound onramp with the T2 lanes they completed.
I’m really surprised (and disappointed) that the new bus lane along Te Atatu Rd only saves 3 minutes. I would have expected it to be much, much more than that. Great analysis though. I hope AT perform the same types of analysis to see where they can shave times off the routes. The comparison with the Northern Expressway was great too – it gives us a glimpse of how things might have been if they’d added a bus lane to the NW motorway when they worked on it a few years back.
What it doesn’t show is the increased reliability of the service. That’s particularly visible in the afternoon peak, when the T2 doesn’t operate (even thought it should, as the traffic is stationary from McLead’s rd intersection) – the 13 tends to get very bursty from Te Atatū Interchange onwards.
Never forget the genius that was in charge of the NW motorway expansion that shafted PT users with nothing but hopeless occasional buslanes, and further anti-futureproofed it for any kind of rapid transit, was then put in charge of light rail.
On what planet would stuffing up one major route for public transport earn you the job running the biggest next one? And look at the great and predictable disaster that was. How good are NZTA at killing PT? And how hopeless was Labour for letting them?
NZTA didn’t kill Light Rail in NW Auckland. Close to SFA happened with regards to the North West route. Ultimately it was just not even enough a priority to do much work on it, as the OIAs that were declined due to the documents never existing in the first place can attest. That’s an entirely political decision around delegation, and nothing else. You can’t blame the agencies if they were never seriously instructed to actually progress it.
Was it never a budget line?
Exactly – NZTA was not tasked with building rapid transit to NW Auckland. The fault lies with the National Government for not progressing it during their term, and then Labour for also not progressing it urgently either.
What a great analysis, very interesting – well done!
The route is being designated/protected thankfully. Whether it’ll be a busway ‘future proofed’ for a conversion that’ll never come, or a better mode honestly depends more on the private sector. The gov will prefer to plonk down some tarmac and run buses, private sector will do something higher order.
Was meant to be a reply to BW about NW from Westgate to Kumeu.
yes
Wow, this is amazing work, thank you. I travel from Westgate to K Rd and your travel time graphs are really useful – wish AT would provide this kind of info!
Before the changes, I used to catch the 125x which would travel along Gt North Rd, but unlike the 11T/W it did not stop. I wonder if that is an option for the WX1? It certainly felt faster than the motorway does now, but that could just be rose-tinted glasses.
And agree the likelihood of getting left behind seems small. It’s only happened to me once.
The WX1 is certainly a marvel, when you get it at the right time on the right day, making a trip half the time and not much more than a car especially considering there are no parking concerns. However, there are a few problems:
1. K rd.
Like others the trip off the motorway along KRd when the bus lanes aren’t operational are painful and it should be changed to much longer hours during the day. There are several side roads people can park on instead.
2. The bus lanes and on ramps. Having been on the bus after the designated ‘BUS LANE’ hours when the motoway or onramps have been out of hours and therefore compelled to wait in massive queues is beyond frustrating. The whole benefit of being on a bus is to whiz past cars stuck in traffic. This seems to be a problem that AT has with NZTA and it needs to be resolved asap. Its just stupid.
3. While it’s better getting to the city it is pretty much always worse now getting to Pt Chev/Western Springs. Having to get off and wait up to 20 minutes for an 11, which happens to me frequently has added 30 minutes to my commute as this is where i work . The only solution if for the 11 buses to be more frequent.
4. The new T2 lane on Te Atatu rd is great but the hours need to be expanded particularly for after school. It can take 40 minutes to get from Henderson to the interchange at Te Atatu and as it can also take 40 minutes to get up Lincoln Rd to the Lincoln Rd Interchange at anytime from 3pm onwards . Either way you’re in for a long ride and then if then another 20 minutes because you missed the 11 by 30 seconds.. aaah… very frustrating !!
4. It’s a start , let’s hope they keep on with it, better bus stops, more buses at night , more T2 lanes.
Yes, the original proposal had afternoon operating ours on Te Atatu Rd
https://at.govt.nz/projects-roadworks/te-atatu-t2-lanes#outcome