Share with public all data on bus service standards

Clementi is the worst place to start from when going downtown, as I recall from a news story a month or two ago. Tampines also figured in the hellish-commute stakes. A figure of 20 minutes was mentioned, increasing to a little over 30 minutes at peak hour, if my recollection’s any good.

The times sounded too good to be true — 30 minutes is hell? —  and I did a double-take. Only on re-reading the article did I realise it was about driving. The study did not refer to the proletariat that had to rely on public transport.

It only upset me more.

Heck, some days, I am kept waiting 25 minutes just for a bus. You may have reached your destination, but I am still stuck at my starting point. It happened again over the Christmas weekend. It was raining and I had to rely on a feeder bus, which took forever to arrive. I resolved to write.

The service quality standards laid down by the Public Transport Council (PTC) can be seen at their website In a nutshell, these are the standards required of bus operators:

“Headway” means the time between one bus and another on the same route. However, the service standards only refer to headway at the commencing bus terminus. The 10-minute or 20-minute headway intervals mentioned above (with 5 more minutes’ allowance) do NOT apply to buses en-route.

Nor could I find a definition for “peak” and “non-peak”.  I would have thought it essential to set out the times clearly if the quality standards are to mean anything.

Just today, I see a report about SMRT Corp, the operator of the East-West and North-South metro lines, increasing train frequency:

Those travelling at the peak of the weekday morning rush hour will see trains arriving every 2.14 minutes to 2.5 minutes, while trains will arrive every 2.5 minutes to 3 minutes at the height of evening peak hours.

The morning peak hour stretches from 7am to 9am, and evening peak hour starts from 5pm and ends at 7.30pm.

— Sunday Times, 1 Jan 2012, More frequent train service during rush hour, says SMRT, by Royston Sim.

Many readers would leave with the impression that they can expect a frequency of 2.14 to 2.5 minutes during the morning peak of 7 to 9 a.m., but you would have been fooled. Read it again. It says: “at the height of peak hours”. What does that mean?

It is also rather irresponsible for the newspaper to place a sentence about peak period timings right after SMRT’s claim which DOES NOT refer to peak period timings.

Coming back to buses, commuters may need to judge if the standards are too lax. Is a 10-minute headway (ex-terminus) during the mysteriously undefined “peak” good enough? Is a 20-minute headway all other times good enough?

Of course, these are minimum standards. The bus operators SBS Transit and SMRT could well be doing much better than that for several services, but it would be hard for the public to know, since this is not published information.

More importantly, what matters to commuters is the headway en-route. We hear of buses bunching up followed by a terribly long interval before the subsequent bus comes along. The latter will likely be overcrowded as a result. In this respect the Public Transport Council has no quality standards in operation. It appears that the PTC takes the bus companies at their word that it all depends on traffic conditions and there is nothing they can do about it.

Worse, I haven’t seen any reports about attempts to measure headways en-route. It’s a known unknown, but we make no effort to find out?

* * * * *

This is where I beat the drums called Freedom of Information, and the related Open Data. If the PTC won’t act, let citizens act.


All our buses use (or should be using) the Global Positioning System (GPS) or other systems to predict and record arrival times at bus stops. At several bus stops, we have electronic boards like the one at right.

You can also digitally query SBS Transit about the expected arrival time of the next bus of any of its services, at any bus stop.

Clearly, data is being collected. Release the raw data (in machine-readable format) to the public and some whiz-kid somewhere is going to find a way to crunch the numbers to reveal patterns and trends. We’ll be able to see how bad the problem is: Which routes tend to have the problem of bunching and/or long headways en-route, and at what times.

We can then focus on possible solutions such as looking at more efficient bus lanes along the most troublesome stretches of roads, or redesigning the service routes if their excessive length aggravates the delays.

SBS Transit and SMRT may well say, Oh, we’re already doing that. Maybe they are, maybe they aren’t. We, the public, don’t know. And that is the point. Since they are running a public service, we should have a right to know. We should have a right to monitor their performance.

In theory, the PTC is supposed to represent the public in monitoring the operators, but given the complete lack of information from the PTC — “peak” not defined? — and the apparent failure to even measure en-route headways, it may well be sleeping on the job. In any case, as a public body, the PTC too should be sharing information with the public. There shouldn’t be a cosy, opaque relationship between the regulator and the regulatees.

The time for open data is now.

16 Responses to “Share with public all data on bus service standards”

  1. 1 SCL 2 January 2012 at 12:56

    Officials must walk the talk.

    To truly do their job, they must make use of the services they manage or provide.

    Lui and Ee must take public transport every week. Saw must take the MRT to/from work and other appointments numerous times weekly, etc, etc, etc.

    Otherwise, how can they know what they are truly delivering? Not everything can be measured. Some must be felt and experienced, and frequently.

    In all other businesses, managers meet their customers and get direct feedback. Why not government and public transportation?

  2. 2 Vote for Change 2 January 2012 at 16:52

    Yes. I totally agree with you on the need for PTC to wake up it’s idea. But even the online data is not accurate. Some times the timing shows fake arrivals: buses that do not come. I wonder how on earth the buses are tracked if they are actually tracked at all.

    • 3 yawningbread 2 January 2012 at 23:07

      Fake arrivals I have experienced, but to be fair to SBS, my observation is that 90 – 95% of the time, the data is accurate. However, I don’t use the system regularly, so I can’t say how reliable is my experience.

      • 4 Vote for Change 3 January 2012 at 01:25

        Based on my personal experience, I would also put the data accuracy at between 90-95%. The idea of having an estimated arrival time is good. But the way the system is implemented is deplorable.

        How does SBS tracks a phantom bus with GPS complete with arrival timing estimates? Is LTA using the data provided by SBS to assess their quality of service such that SBS is trying to game the system?

        LTA and PTC doesn’t seem to mind that the information provided by the transport operators is 99.999% accurate (please note, precision and accuracy are two different meaning). Or do they even check at all? What is the point of investing in such a crappy system which provides fake information at times? Is public money involved?

  3. 5 Dee 2 January 2012 at 21:41

    The only way to really check for bus timings is through some app called SBS Next Bus. Based on hearsay and personal experience, it’s a big improvement over SBS’ Iris system which is … kind of rigid and a pain to use? SBS Next Bus is available for Android, iPhone and so on but it’s not always that great. My grouse is that it only accepts numbers and not alphanumerical names like “Number 5” while entering “5” yields an error the bus service isn’t recognised. And it also crashes like Windows 98 used to.

    Furthermore, it’s a bit silly that we’ve to rely on a smartphone app, for the bus arrival time at a certain bus stop. What about those who don’t have a smartphone? Or rather, what if your smartphone is out of battery and you need to go home urgently? And it’s also a bit harsh to tell someone to take a taxi since that’s very expensive!

    • 6 Dee 2 January 2012 at 22:40

      Also, 30 minutes is not hell. When commuting to work, a family member had to wait for over an hour just for the bus to arrive. Sometimes she opted to take another bus number instead and then switch, in hopes it’d be faster. And at times, the waiting time would be from 2 to 3 hours ‘cos there was no bus arriving at the bus stop.

      So much for efficiency.

  4. 7 KC 2 January 2012 at 22:08

    Policy makers could spend three months going everywhere in public transport, make that months of Feb, May and Dec.

  5. 8 anon4cec 3 January 2012 at 10:19

    1. Ghost arrivals happen often enough, so one must assume there’s an override programmed into the sytem.
    2. Bunching happens often enough, so one can be sure the regulators are still time-warped using operating metrics designed for kampung days when computers used punch cards.
    3. We don’t even have a holistic sysem-wide smartphone app that covers all transport operators.
    This rojak buffet of bad outcomes is all in the name of competition. The technology’s there. It’s abused where it suits. It’s not used where it should. Don’t expect transport companies to do anything about it. It’s the regulator’s job. Hellooooo??!! are you there?@$^&!##!!!

  6. 9 georgia tong 3 January 2012 at 12:46

    Yes agree, the data displayed at the Bus Arrival board is not accurate. It flashes ‘Arrived’ but no bus in sight for at least another 5 to 10 minutes – that is if it arrives at all. For those who have just reached the bus stop – they thought they missed the bus when in fact it is the bus that failed to turn up. Another more serious issue we have to content with is unsafe driving habits of many bus drivers. The moved off even if elderly passengers are struggling to make their way to the seat. Can’t they just wait a little longer ? They brake so hard that even those seated slide off the seats ! I have seen children falling and even adults struggling to maintain balanced. It is not safe for elderly to take bus on their own anymore nowadays.

  7. 10 honcheng 3 January 2012 at 23:09

    Hi, good post.
    I’m a mobile developer for one of the free bus apps in Singapore. Not sure why information on public transport, is not public data. There are lots of developers like me that will be interested to make mobile apps for free, but they can’t because SBS requires a signed agreement with them, and SMRT doesn’t have an official feed for arrival time.

    I agree that SBS arrival time data is more than 90% accurate, and that is sufficient. SMRT has ‘protected’ access to it’s arrival time feed on their website, that isn’t accurate most of them time where I use them. Sometimes they are off by a few minutes. At Clementi station, usually the timing jumps from a 4-5 minutes, to arriving suddenly. Not sure if they are mixing up schedules with estimated arrival time based on GPS.

  8. 11 Eric 5 January 2012 at 09:48

    It’s also not clear whether the data is collected and actually reside in a database to be mined or published. The information could also be transient, or purged at a fairly regular basis.

    But i wholeheartedly agree with you about implementing a similar freedom of information act as the U.S. government, especially for demographic data. Singapore conducts census regularly, but I get the sense that none of the raw data is available for finer-grain analysis beyond people working in or for the government.

  9. 12 Chanel 6 January 2012 at 15:39


    The PAP govt is not going to give us a Freedom of Information Act and in the remote chance that they did, it would be so watered down as to be ineffectual.

    For political party perpetuity, this govt wants the public to only access “information” that has been massaged to project a good image of the govt.

  10. 13 MEMorris 11 January 2012 at 12:31

    I fully agree that a ‘freedom of information’ policy for any & all data related to tax-payer funded programs, initiatives, etc is crucial to foster a greater sense of self-determination of Singaporeans.

    Companies like Google are continuing to contribute to better understanding of real-world activities/trends and how they are reflected in on-line searches via their “Google Correlate” tool, which is available now (and even traverses Google-Singapore searches as an option). You can even upload your own data and have Google-correlate map its database of searches against you these data, looking for correlations between the search-world and your dataset. Tools such as this can help to illuminate the public’s thinking and could be helpful in detecting trends/events if properly used. Here is a link for those who might be interested:

  11. 14 usherer 16 January 2012 at 00:15

    In other countries I’ve lived in, there are buses of different sizes and quite a number of them would cover similar routes so you have many options. Over the last 30 years, the bus routes im familiar with have either not changed AT ALL, or have become worse with fewer options. I dont drive,i have limited earnings – which means i hvaent visited many parts of my own homeland, or have only been there once or twice in my lifetime. and im a fully able-bodied person in a 600 sq km country. what about the old folks / disabled/ teenagers? why are we held captive in our own home? I read with much disappointment that the farms are no longer able to keep on hiring bus service…

  12. 15 nhy 25 January 2012 at 22:08

    Both SBS Transit and SMRT websites provide the bus arrival time. However, they require a verification code — presumably to defeat bots — and only provide one bus service at a time.

    LTA’s MyTransport.SG portal is better. It provides all the buses arrival times at a given bus stop.

    The given information is limited. Short of polling each bus stop at a high frequency (say 15s), the only reliable information we can gather is the interval between two buses.

    This should be good enough to determine the peak/off-peak frequency — if the data is somewhat reliable.

    The portal only has information for 212 bus stops. If we combine the timings from all the bus stops, we can get data for around 172 bus services.

    For some buses, it is possible to map the bus route to see how the bus frequency changes over the route. However, this is quite sparse (given that there are only 212 bus stops in total.)

  13. 16 nhy 30 January 2012 at 12:58

    I have two questions regarding LTA’s MyTransport.SG portal. I don’t know where to post them, so I might well ask them here. 🙂

    1. Sometimes the next bus is earlier than the arriving bus.

    2. Sometimes, a bus that shouldn’t appear at a particular bus stop shows up.

    Has anyone else observed these?

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: