Paratransit in South Africa

[Johan Joubert]

Hi, unfortunately I was not in-the-know yet about this group when you had your workshop last year. Hope to catch up soon.

In South Africa we also have paratransit in the form of minibus taxis (jitney-like vehicles). Although the popular view (even in government) is that this type of paratransit should be formalised, we have the view that the minibuses are part of our South African solution to the out-of-whack urban form we have, and not the problem, as they are usually portrayed [1,3]. My interest in GTFS-for-the-rest-of-us is mainly two-fold:

1. our paratransit accounts for more than two-thirds of 'public transport' trips. Still, there is a large barrier-to-entry to use the mode, mainly because of a lack of local knowledge. Vehicles are not signed, routes are not documented, people don't know the correct hand-sign to flag down a passing taxi (hand signs are used by the commuters to indicate intended destination, see [2]). Same problem as elsewhere where this type of mode operates. We've done some 'manual' capturing of routes in an attempt to play around with answering "how can we visualise the routes?". Attached is one such an attempt for one of the route associations in Tshwane (Pretoria). The darker the read line, the more prominent those particular road segments are serviced. At least in South Africa, many people believe that the minibuses just drive around anywhere. But they don't. They follow 'fairly repetitive' patterns. The little blue dots on the attached figure is actual waypoints where boarding and alighting was recorded. again, the darker the dots, the more frequently those points were serviced.
2. We use the large-scale Multi-Agent Transport Simulation (MATSim) as a tool for transport modelling and transport planning. MATSim is now capable of modelling paratransit, thanks to the Ph.D. of Andreas Neumann (he'll be defending it soon) from the Technical University Berlin. A paper will also appear shortly [2] covering the South African case. If we can, eventually, capture the GTFS data for the flexible mode, then we can use it as a valuable form of validation for our behavioural models. We believe that unless these flexible modes, which accounts for a large proportion of commuters, are not accounted for in the transport planning models, such models render quite mediocre decision-support.

I hope I'll be able to contribute to this group. What have we done so far? We've started a few years ago, using Trillium's TransitDataFeeder to capture the first few GTFS feeds for a South African rail operator, Metrorail. Since then things were quite slow, but recently we got Conveyal's Transit Data Manager up and running. Still playing around on how to use our available minibus data and get it in there (the attached figure was simply using R). It would be interesting for me to know how the other groups are doing this.

Regards,

Johan

References:

[1] Joubert, J.W. (2013). Gauteng: Paratransit - perpetual pain or potent potential? Chapter 6, Megacity Mobility Culture, 107-126. Institute for Mobility Research (ifmo) (Eds.). Lecture Notes in Mobility, Springer Berlin Heidelberg.

[2] Neumann, A., Roeder, D., Joubert, J.W. Towards a simulation of minibuses in South Africa. Accepted for Journal of Transport and Land Use. Also, working paper 14-03 from the VSP group.

[3] Woolf, S.E., Joubert, J.W. (2013). A people-centred view on paratransit in South Africa. Cities, 35, 284-293.

[Neil Taylor]

Hi Johan

Great to hear from you, and welcome to the group!  It's a real shame you weren't aware of us earlier, but work continues to build a resource center that will enable people at transit agencies, transit operators and government agencies from around the world to more easily understand the GTFS spec and use it to model their public transport services.

It is really good to learn that you have been using Trillium's Transit Data Feeder (Aaron Antrim is part of this group) as well as Conveyal's Transit Data Manager (which I'm proud to have helped develop alongside Kevin Webb and our colleagues). The two areas of interest you have identified -- 1) improved information on public transport, and 2) improved transport planing decision-making -- are fundamentally the same as those other members of the group have identified.  A common third is the challenge of persuading government officials and decision-makers of the value they stand to gain from investing in open transport data.

Re: point 1 in your message, this is the primary issue for many of the city authorities I've met (and many of the members of this group).  We've also found that in most locations the paratransit services are largely predicatable and well-known to local transit users.  Our efforts have been to try and influence the GTFS spec so that it incorporates modest changes that make it easier to model paratransit services such as the minibus taxis in South Africa.  These efforts are documented in this group, and in the GTFS Changes discussion group.  The only thing we need to effect this change fully is a small number of cities/operators prepared to encode their data based on the minor changes proposed by the GTFS-for-the-rest-of-the-world group (allowing for 'non-stop-based' services to be encoded, and for colloquial vehicle types - e.g. 'Jeepney' or 'Matatu'- to be included in public information feeds).

Re: point 2 in your post.  Work we've been doing with colleagues at SETRAVI in Mexico City D.F. focused initially on developing public information feeds in GTFS, but is increasingly focused on developing open source public transport and city planning tools (for accessibility analysis to inform better decision-making) which is quite similar to the outcomes you are trying to achieve using MATSim.  Interestingly, and possibly for discussion offline, my colleagues at ITP have created multi-modal transport models for Manila and Cebu which incorporate the movements of buses and jeepneys using Cube, but this has largely been incidental to GTFS, so is largely off-topic for this group.

After the initial group goals of developing internationally-relevant learning resources, and influencing the GTFS spec to make it more relevant in cities where a high proportion of transit riders are using paratransit services, there is plenty to think about.  One such future problem is how the GTFS Real-Time spec could be employed to better understand, predict and communicate the variance that is inherent in a number of the paratransit services found in developing and transitional regions.

I hope that is a useful introduction to the group, and look forward to chatting more

Best regards

Neil

[Johan Joubert]

Thanks Neil.

You mentioned "[t]he only thing we need to effect this change fully is a small number of cities/operators prepared to encode their data based on the minor changes proposed by the GTFS-for-the-rest-of-the-world group"

We have the "will", i.e. I will take responsibility and commit to do that for the (bit of) South African data we have, mainly in two cities: Tshwane (Pretoria) we have a few routes, and Pietermaritzburg (here we currently have one). Maybe just to start off with. There's more data, but I cannot bargain/negotiate for them if I cannot show (other than the visualisations provided earlier) what we're doing with the data. I just don't know HOW. Is there somebody specifically that I can talk to (maybe offline). Have somebody got a separate branch/uncommitted version of Trillium or Conveyal's tools that have already been adapted to capture the proposed changes?

Thanks,

Johan

[James Smith]

Hi Johan,

Not sure if you will find something that meets your needs, but have you seen this repository which includes links to a variety of GTFS related materials?

https://github.com/WorldBank-Transport/GTFS-Training-Materials/wiki/Link-repository-for-international-GTFS-training-materials

cheers,

James

James Smith

REEEP

--
You received this message because you are subscribed to the Google Groups "Making GTFS Work for the Rest of the World" group.
To unsubscribe from this group and stop receiving emails from it, send an email to making-gtfs-work-for-the-re...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Ahmed Dorghamy]

Dear Johan,

I just joined this group because I'm doing relevant research on Egypt. It was very interesting and informative to read your e-mail, thank you!

I am just starting my PHD at Humboldt University on the impact of informal transportation on travel behavior in Greater Cairo, with a case study of a prospective BRT corridor.

I'm in Berlin (Prenzlauerberg) I'd love to get in touch and exchange ideas if you're interested, especially that I'm planning to do some mapping of paratransit in Cairo and I'd love to know more about your experience.

Best regards,

Ahmed El-Dorghamy

Mob: 0152 070 23667

ador...@cedare.int

[Johan Joubert]

Ahmed,

thank you. You're welcome to contact me via Skype (JohanWJoubert). I will be in the Berlin area  (or at least in the vicinity) end August/early September.

Regards,

Johan

PS: One of my co-authors, Andreas Neumann (paper [2] in my original post) has since completed his PhD (at TU Berlin), which is available from here.

[Ofentse Mokwena]

Hi All,

I recently grappled with the publication in the JTLU and then followed through on Andreas' PhD effort. I'm barely clear on GTFS, so I'll spend some time familiarising myself with the effort. I'm a lecturer at the North West University, South Africa, and concluding my Masters with the University of Cape Town. A few weeks ago I submitted pilot study results for a stated preference study for minibus and bus operators. I'm intrigued about integrating utility functions into the routing/mapping and choice behaivour of various operators-- especially at terminals (a potential PhD theme).

This group appears to be an exciting effort.

Regards,

[Jackie Klopp]

Hi Oftense, 

Great you are interested in this work. Let us know what you are interested in doing. 

You might be interested in the work around Cape Town by Andrew Kerr <taxi...@gmail.com> see also  www.taximap.co.za. He has not put his data into GTFS yet but has been collecting data on routes and where they go.

Welcome to the group!

Jackie

--

You received this message because you are subscribed to the Google Groups "Making GTFS Work for the Rest of the World" group.
To unsubscribe from this group and stop receiving emails from it, send an email to making-gtfs-work-for-the-re...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Jacqueline Klopp
Associate Research Scholar
Center for Sustainable Urban Development
Earth Institute, Columbia University
475 Riverside Dr.
Suite 520
New York, NY 10115
(tel) 212-851-2979
(fax) 212-851-1780
skype jmklopp