I've been poking around with the fares data available here:
http://data.atoc.org/fares-data
(by converting the text files into a postgres db)
And wondered if anyone here was familiar enough with it to tell me where I'm going wrong.
What I'm trying to get to is something where I can spit out, e.g. the cheapest (non-advance) return ticket from a given station, and I think I'm successfully connecting stations via flows and non-derivable fares using station clusters where appropriate - still ending up with some gaps though, e.g. my output for Chichester (CCH) gives the following 'missing' stations:
SAN | Sandown
SHN | Shanklin
BDN | Brading
CCH | Chichester
HOH | Harrow-on-the-Hill
LKE | Lake
RYR | Ryde St Johns Road
SAB | Smallbrook Junction
Now, obviously makes sense that there wouldn't be a fare from a station to itself, so I can ignore that.
Harrow I'm guessing maybe I could find using the group locations from the Locations - i.e. there's probably a fare from Chichester to "London Zone 1-6" which Harrow seems to be grouped into. Although why it would be the only London station I need to do that for I don't know - is that likely or would it suggest I've screwed my logic up somewhere?
Where I'm really stumped is the other stations, which are all on the Isle of Wight - they're all part of a cluster which I can connect to various other clusters, but none of which appear to contain Chichester (and there don't appear to be any useful group locations).
Am I missing some other grouping/clustering rule that would help me join the dots here?
Thanks in advance for any help!
http://data.atoc.org/fares-data
(by converting the text files into a postgres db)
And wondered if anyone here was familiar enough with it to tell me where I'm going wrong.
What I'm trying to get to is something where I can spit out, e.g. the cheapest (non-advance) return ticket from a given station, and I think I'm successfully connecting stations via flows and non-derivable fares using station clusters where appropriate - still ending up with some gaps though, e.g. my output for Chichester (CCH) gives the following 'missing' stations:
SAN | Sandown
SHN | Shanklin
BDN | Brading
CCH | Chichester
HOH | Harrow-on-the-Hill
LKE | Lake
RYR | Ryde St Johns Road
SAB | Smallbrook Junction
Now, obviously makes sense that there wouldn't be a fare from a station to itself, so I can ignore that.
Harrow I'm guessing maybe I could find using the group locations from the Locations - i.e. there's probably a fare from Chichester to "London Zone 1-6" which Harrow seems to be grouped into. Although why it would be the only London station I need to do that for I don't know - is that likely or would it suggest I've screwed my logic up somewhere?
Where I'm really stumped is the other stations, which are all on the Isle of Wight - they're all part of a cluster which I can connect to various other clusters, but none of which appear to contain Chichester (and there don't appear to be any useful group locations).
Am I missing some other grouping/clustering rule that would help me join the dots here?
Thanks in advance for any help!
