Source data problem?

[dbrb2]

Hi,
Looking at the Lithuanian data, a number of the stations appear to be towns with no railway station - for instance:

Prienai
Lat: 54.632545
Lon: 23.94787

Looking at a map though, I don't think the town has a railway station...presumably this is down to the source data?

[misdoro]

We do sell bus carriers as well, plus multiple cities have more than one station.

For those two reasons this database includes the cities and towns in the covered areas, either as meta-stations covering multiple railway stations, or simply as stations to be able to search for the bus operated services.

[OlgaDezh]

@misdoro , @lucavezz could you advice how to detect whether the record describes a bus stop or a railway station?

[misdoro]

I'm afraid there is no reliable way to do that, we do not have such information.
One approach could be to check if the station has only busbud_id / distribusion_id / flixbus_id values defined, however this is not a reliable approach. For ex in Switzerland there will be multiple bus stops that are available through CFF.

[OlgaDezh]

@misdoro apart from CFF, is there other carriers among the ones listed in the lookup that provide both bus and rail service?

[misdoro]

SNCF has some bus services as well, other carriers may have them too. So if you want to exclude bus stops there is no reliable way to do so.

[OlgaDezh]

@misdoro , thank you

[Mark531]

So if you want to exclude bus stops there is no reliable way to do so.

In other words, this dataset is useless...

[fbarbe00]

Correct me if I'm wrong, but wouldn't removing all rows where uic is None only keep railway stations?

[Mark531]

So, I filtered the dataset on Switzerland and UIC not null: I get 500 stations, far from the expected ~1600. As for France, it's the opposite, you come up with many more stations than expected (6,518 vs ~4,000).