Crowdsourcing: BMTC Vajra Route info database - I

 Now you can see the in-progress Google Woksheet within this page. Update is progressing slow, but steady. 

 Check this link    and today's news article.  

.....

“We believe that this initiative will be highly useful for the millions of people who use BMTC buses for their travel and daily commute. We are proud to be the first bus service to provide this kind of service on Google Maps to its commuters. The collaboration with Google is another step in this direction to make it even easier for the public to plan their trips with transit information on Google Maps”, said Syed Zameer Paasha, managing director, Bengaluru Metropolitan Transport Corporation (BMTC).

Manik Gupta, head of Maps and Local Products, Google India said, “We are extremely happy to collaborate with BMTC to offer this service to commuters in Bengaluru. Google Transit's goal is to make public transport information easily accessible and useful, and we are delighted to extend our Transit offering to include buses.”

He added that they have on offering the users a easy way to get all the information they need to plan their trips as per their convenience. Commuters can get all the information like showing bus frequency, route, and travel duration.

Bheema/Arun,

1. Are we 'claiming' copyright/ownership of the data? If not, we should be fine.

2. BMTC has only introduced these routes. But, it is not that BMTC 'owns' the information about the route. This knowledge is already in the public domain. If I say that Manmohan Singh is the prime minister of this country, nobody (including GOI) can claim ownership to such a piece of information.

3. In my opinion, we should not bother to host the feed nor should we deal with Google ourselves. BMTC is already working on a Google Transit feed (by the end of the month, it should be having the feed with info for the branded services - Big10, Big Circle, Vajra and Vaju Vajra). We just have to pass on this information, labelling it as crowdsourced info, to BMTC and let them deal with Google/thirdparties. Let us not 'hope' to play a role that can be percieved as competing with BMTC.

Also, why would Google host or rely upon data given by  Praja. Look at what Google writes at http://maps.google.com/help/maps/transit/partners/participate.html

If you provide a transportation service that is open the the public, and operates with fixed schedules and routes, we welcome your participation - it is simple and free

So, let the responsibilty of data (and 'ownership') remain with BMTC!!

  If yes, please let me know the phrases to be displayed there to protect crowdsourced data from any legal stuff

Content is being provided in public interest from various sources whose accuracy Praja.in has not verified. The data is being provided on an as-is where-is basis with no warranties implied or otherwise. Praja.in is not responsible in any way for any consequential damages or liabilities arising out of subsequent usage of the data.

Hi bheema, i just noticed that i had posted #busroutes.in instead of #busroutesin

I was really hoping to see you, sorry for the goofup. The chat logs are here. We'll have another meet in a couple of days, i'll let you know.

I will try to join the meet, but not sure. Meanwhile I am continuing to focus on data gathering in manual mode, though I am aware its very HUGE amount of work collect data manually. I appreciate your effort to converge the efforts into gtfs feed from busroutes.in.  I am sure arun and team will have this link, but just to mention, here is the link to gtfs specification. Proably from there, you guys can pick up the thread.

Shastri,

I have read in few forums that Google guys take few months sometimes upto a year to open up Google maps even for preview mode. I heard even New York city not opened Google Maps. So is it the good time to start  to communicate to Google?

Bheema, me and a bunch of others who have put together http://busroutes.in are have an IRC meet tomorrow, tuesday night 10pm.

The plan is to figure out the way ahead to rewrite the app to help crowdsource and geocode transit data to create a GTFS feed. As you have a good knowledge of this format you can share your expertise as well as ideas that you have for the bangalore gtfs feed.

Of course, the meet is open to anyone who is interested in joining the discussion

IRC: freenode#busroutes.in

Busroutes.in google group

 Hi Sanjay,

I appreciate your efforts. However this question raised earlier by ss87. Here is the answer:

Buses may stop in different places for example : Garvebhavipalya. The stops for Hosur Road towards Hosur(South) and towards silk board(North) are at least 50 meters away. So ensure seperate entries for them. Example : Garvebhaipalya(N) and Garvebhavipalya(S) with distinct stop codes. Append new row with new data in stops.txt.  Another way is to add direction. Example in stops.txt: Richmond Circle(Towards Mayohall) and  Richmond Circle(Towards Corporation)

Further when you create trips, you should specify in column "direction_id" 0 being onward journey and 1 being return journey in same route. This means, you can have different places for same stop in same route. Even return route may be different if some one way restriction imposed. Check for route no.100. Open trips.txt you will find 2 trips with 1000600 and 10006001(check for direction_id column)

Should you need more clarity, feel free to contact me.

 I was talking about computation considerations. I have made speed as input parameter in trips.txt. We can play with the same, be cautious it will change stop timings :). Once you complete playing with speed, revert to original value. Also you can play "signals" in stop_times.txt. You see how the re-compuation done to arrive stop times. 

Good progress on stops.txt !

Could you tell me where you obtained data about the no. of stops (ie.1764 stops) ? As per BMTC route booklet (slightly outdated), there are only some 1200-odd, but the data does not include many areas outside ORR.

 FAQs updated.

13.Tell me about distance calculations and travel time estimations "As we do not have accurate arrival/departure timing available , we are estimating based on distance between stops. The distance formula is taken from www.movable-type.co.uk/scripts/latlong.html  Thanks to that author. Now you bangalore local conditions will not match with direct speed calculations. So have implemented few factors which add to travel time.

a) In trips.txt u will see few columns in last => average speed, peak days(not used) and peak hours(not used).  

b) In stop_times.txt you will see new columns added to right most. est_distance,signals and est_traveltime."

Adugodi 12.94374134 77.60809779

Adugodi Pol Stan 12.93916147 77.61021137

Distance in KM 0.5583975385843

So now arrive at following computation

est_distance=using formula at cell C19

est_traveltime =est_distance*avg speed for trip+(no of signals*3 minutes). Formula is given below for techies:

=L243/VLOOKUP(A243,trips.txt!$C$1:$K$2501,7,)+M243*(3/60)

(est) time of arrival= departure time of last stop + est_travel time

(est)time of departure= arrival time+ 2 mins

Note: No computation for trains as we have time tables

Any comments and inputs regarding speed calculation factors?

I particularly quote the below observation from that link which is the crux of Route rationalization we have been harping on in these forums for a while now

 ...entire city can be covered with very few Trunk Routes and hence the number of trunk routes that is needed will be much less than the current system. In the case of Bangalore, BMTC is yet to implement Grid Based services on full swing. But they say that they have identified 15 Trunk Routes for the entire city, which is much better than current 1700 randomly oriented bus route system in the city.

Complimentory reading:  A  paper regarding  BMTC bus route advisory algorithm  from a PES student. A good read.

Planemad,

Thanks for mentioning http://busroutes.in

In fact, Bheema and I had an offline discussion about using Busroutes.in when we started doing this.

We concluded that busroutes.in presented a wonderful way to crowdsource the data. We considered hosting an instance of busroutes on praja.in itself and carry on from there.

The biggest problem with that approach was that we do not believe we could realistically beat the popularity of Google maps among Indian users. Hence we just went ahead and decided to stick with GTFS.

So busroutes.in presents the best way to collect the data and GTFS remains a way to share it with the world.Since we have just begun, I believe there might be a combination strategy that uses both these strengths.

If we could come up with a busroutes.in to GTFS translator, we can collect all the data on busroute.in directly (or on an instance on Praja.in server) and prepare GTFS bundles for Google to pick up.

That way we keep the route info up-to-date through crowdsourcing without depending on BMTC or BMRC and share it on Google where most people can use it.

Do you think it is easy to write a translation tool? Any python wizards on praja who can do this?

Shastri

 500 stops geocoded, 15 BIAS trips added, Almost all trains covered ! Slow but steady updates.. Keep watching and contributing..

 By the way anyone has idea about private buses and other shuttles who runs services systematically? We can add them too, we are not reserved for Govt, services. Dont you agree with me? I think 75% accurate data will do.

• 300 of 1764 stops are geocoded . I would request local people to update unfilled geocodes for bus stops near to you. Click here to start/continue contributing.
• 2175 routes added
• 9 Local Train trips data added to begin with.
• Timing data for each stop added for 33 stop points.

Its just beginning. Long way to go...

 @idontspam, thanks for info. I will update those few services as local trains for now and let us update data when local train dream becomes true one day. I will search data at NammaRailu blogs. 

 235 of 1764 stops are geocoded with exceptions. I would request local people to update unfilled geocodes. The left out stops could not be found on google maps. Hence only local guys can help there.

Another Good news : I have added Namma Metro, Existing Commuter Rail System(South Western Railway local trains) as local transit agencies. This means we are providing integrated data to Google Maps. In other words, when Google Transit for Bangalore is launched it should suggest an all possible ways to public travel in Bangalore to a Googler.

Request: We need full time table of local train services which cover all 34 local stations as mentioned in Praja somewhere.

And Prajas, keep updating https://spreadsheets.goog... as you get time..

Here the authentication link to Google Transit Doc. Hope this would work for everyone. My apologies.

https://spreadsheets.goog...

 Submssion: I am new to Google Docs.

Hence experiments with access levels too. Yesterday I saw people accessing it but it showed anonymous. Now again made it public so that all people can edit without sign-in, though it will show anonymous. But there is a column called comment. Request people to put their names and comments there.

@silkboard: Can you try again. Hope this time it should be ok.

@idontspam: It is too early for prototype. For a working prototype all sections of that worksheet has to be updated with relavant date. Then we have to host process data as a feed to Google. Then we have to contact Google to pick data from feed. Once they(Google) are ready, they will give a preview link. Let me know if you further clarification. Sorry for misleading you, I was just mentioning that I completed 140 routes data to a specific section("routes.txt").  It was just to promote, so that public take more interest :)

Update:  150 of 1764 records updated in "stops.txt"

Hi Shastry,

Will check and respond...

-Bheema

 Good Going Bheema! Do you have a sample of how it looks with the current 140 stops?Curious to see a prototype. 

 Hi ss87,

As per GTFS route description should be short description. Example : Via KR Market, Peak Hour Service etc... You(Anyone) are welcome to update the info. Regarding frequency details  please refer to "stoptimes.txt" section of the goole doc spreadsheet  which I have been updating.

Also, I am planning to add columns to trips.txt. They are "no. of signals" and "peak times". These columns should help us to estimate for reduction in average speed of bus. But no solution of traffic jam caused delays :(.

 I have been updating this  link to create a database which will become feed for google transit. I have updated stop and routes from narasimhadatta's website, and want to give full credit to him for collecting the bus stop names and route numbers. Now same is available in the link above. Now we have to go further. For each a stop we need gecode(latitude/logitude value). May I request the interested people to update "stops.txt" section with geocodes, which are familiar. Also people can participate to validate the geocodes by confirming it in last column. Geocodes can be grabbed from any map application like Google Map or Google Earth.

Update: Anyone can edit now. no restrictions.