Hi All,
By traveling and booking tickets in Indian Railways (IR) for so many years now, I have never understood how we can be sure whether the wait-listed ticket we are booking, will be confirmed on the date of travel. If by some way we can gather large amount of data, we can predict whether the ticket will be confirmed. What do you think
***
Hemant Pulijala
Create an app,I think thats going to be a huge success 🙂
Murtaza Amin
What is algorithm for that app?
how will you find the probability?
Bharat Agrawal
Thanks Murtaza …you got me thinking…..
By comparing the query with the historical data.
Result will be displayed in a pattern in which google shows the results
By same train, same time, same stations but different trains, same train different time.
For example,
WL/61 from HYD to DEL Train no 17335 Date of booking (DOB) :07/10 date of travel (DOT) :07/12 – how many times ticket got cleared
WL/61 from HYD to DEL Train no 17335 Dob:NE dot:NE – how many times ticket got cleared
WL/61 from HYD to DEL Train no: NE Dob:07/10 dot:07/12 – how many times ticket got cleared
WL/61 from HYD to DEL Train no NE Dob:NE dot:NE – how many times ticket got cleared
WL/NE from HYD to DEL Train no 17335 Dob:07/10 dot:07/12 – how many times ticket got cleared
And there could many other variations…
Note that this application or site would be just for information purpose and will not guarantee anything.
Murtaza Amin
I wanted to know what set of mathematical rules you will use, I dont find this easy task. lets meet up and discuss.
are you from Bombay? Im interested.
Bharat Agrawal
I know this is not an easy task. I know IR will not share the data. Whatever data there can be, it has to be collected from public at large.
Mathematically it would be just a percentage of times, the ticket has cleared. Currently as of now, I dont know if there could be some other prediction model. May be you can help me with this.
I am not from Bombay. Currently at Hyderabad.
Want to know more about you. Please share your profile.
Murtaza Amin
we can somehow get the IR data, but I dont think that data is of any use for us as it will just show the times the ticket for any particular route got cleared.
I dotn think this set of maths will work, why dont you do a little research, what are the factors that come in picture when you get a waiting list.
1st is it summers or winters?
2nd is it a weekend
3rd si their any probability of IR arranging for an additional coach for waiting list people. (which by the way accourding to my opinion should be mandatory if the waiting list is more than 100)
4th how long is the trains route.
5th how many stops does the train takes (because the mroe no. of stop the trains takes, the lesser the probability of getting clearance in waiting list, remeber as of now this is jsut a luck game)
6th Im not sure with this whenther our govt. gives you reservation from the next or any other station if the quote from the station I applied for is full or not.
research a bit, are you into maths? try. there has to be a algo. to find this.
Twitter @murtuzi Fb Murtuzi
niraj vishal
Train no-12566 pnr-2335900306 Journy Date-20/6/2012 W/L 4 3AC kanpur to muzaffarpur get it conformed.
niraj vishal
pls ans my Query
neel
Good idea. Have you checked out https://www.bing.com/travel – Bing took over a website called Farecast, which used to predict whether air ticket prices were going to go up or down – and therefore help people make a decision on when to buy the air tickets. You seem to be attempting an analogue for Railways (Instead of when to buy based on price prediction, when to buy based on likelihood of clearance). All the best!
Regards,
Neel
Nishant Agrawal
Hm … ambitious. The logic would be something like this,
A train runs from Station X to Station Y daily. Last time, the person who had booked his tickets 13 days ago was the last one to get a confirmed ticket. Next day, 12, then 14, and so on. The algorithm will find out the average, and improve itself as it learns from more data. Everything can be automated. Innovative.
Vijay Khubchandani
You might also want to get in touch with a Seasoned Agent to get more idea on the Factors involved..
Pankaj Upadhyay
Thanks Neel for Valuable Information
Regards
Pankaj Upadhyay
Alexander Gounder
The nexus between railway babus and travel agents, have always been behind this situation, where you are left guessing… You think railway officials will ever let you use their data to ruin this lucrative business they have established inspite of technology being added to the system.
Bharat Agrawal
Thanks for the response people,
Just for trying railway officials what kind of data should I be looking for.
For example,
Travel Date, Booking Date
Booking Status, Travel Status
Train No
Station To, From
Class
Add something that I missed
As I said before, the advanced factors can be looked later.
Regards,
Bharat
Venudhar Bhatt
It is a wonderful idea for sure. It will be quite complex to be reasonably accurate. But I encourage you to explore it well. I am sure it will be very useful if done right. If you can confirm the availability of data from IRCTC then one must get it for last 2-3 years for select trains and then analyze to start building algorithms or logic and test it out for sometime to refine. Post updates if you pursue and make any progress. All the best.
salil
Hi I have created a website just for pnr prediction. Please give your valuable suggestions
jawwad mobin sajid
and someone did it
Amir Mohd.
I have launched a site http://www.trainman.in which does exactly the same thing. The data has been collected over a large period of time and the prediction accuracy seems good. Please check it out and provide your honest feedback. Thanks.