Every working day (except holidays) I go to work. So I leave home and walk to my local tube station: Wood Green. Then I get a train to Kings Cross where I change the train to Farringdon.
Some days, if Piccadilly Line has delays, I change in Finsbury Park. For some weeks, 1 or 2 years ago, I used to always change to Victoria line but I decided that it’s not worth. Now I change in case it has or if I’m standing up and squeezed in the Piccadilly Line and/or the Victoria Line train is just there.
The TFL webpage allows to download the last month or two months travels using a CSV file (I’ll check for some API, etc.).
I did some quick stats to find the distribution of the time. The most interesting result of that period of time (between 15th February and 5th April, but I was 2 weeks away from London) is:
Duration Times %
00:23 2 7,69%
00:24 3 11,54%
00:25 4 15,38%
00:26 11 42,31%
00:27 19 73,08%
00:28 19 73,08%
00:29 20 76,92%
00:30 22 84,62%
00:31 24 92,31%
00:32 24 92,31%
00:33 25 96,15%
00:34 25 96,15%
00:35 25 96,15%
00:36 25 96,15%
00:37 25 96,15%
00:38 25 96,15%
00:39 26 100,00%
00:40 26 100,00%
Min 00:23
Max 00:39
Avg 00:27
Sample 26
So, during that period:
- I did the travel in 23 minutes 2 twice. Record time!
- Hard to read in that table, but the most common commuting duration is 26 minutes.
- Once it took 39 minutes. Something was very screwed… (and another day it took longer because I used a bus, since Piccadilly Line was not working in my zone)
- I arrive 92% of the times in 31 minutes or less
- I could achieve a good punctuality mark planning 33 minutes of commuting time (96%!). More than that it’s really screwed and not worth planning for a normal day.
This is the final result… the frequency table for so short period of time is not a standard normal distribution as I was expecting, probably if I’d do during a longer period of time would look more standard.
Standard distributions are for natural phenomena. Related to heights, we know that there’s people taller and shorter, but there’s not a minimum. No one will achieve 0m height, also there’s not a theorical maximum.
But your train has a maximal travelling speed that is the normal speed if all is ok. IMHO data is correct and will be similiar with more data.
If someone has R installed:
seconds=c(23,24,25,26,27,28,29,30,31,32,33,34,35,46,47,38,39)
reps=c(2,1,1,7,8,0,1,2,2,0,1,0,0,0,0,0,1)
plot(seconds,reps)