iRail / stations

A list of all the Belgian stations and their properties used within the iRail project
http://irail.be/stations/NMBS
29 stars 20 forks source link

Giving stations a magnitude indication #22

Closed pietercolpaert closed 7 years ago

pietercolpaert commented 9 years ago

Gent Sint Pieters is more important than Gentbrugge. Brussels Schuman is less important than Brussels Central, North or Midi. To optimize user experience, we could use an indication of this importance in our CSV file.

Use case: autocompletion

When autocompleting Brussels, Brussels Luxemburg is the first option. This is one of the smaller stations in Brussels and should appear lower.

Which data should we use?

This importance field can be made on the basis of different parameters:

All of these could are valid indications. We can use these indicators to categorize stations.

Implementation

Add a field importance: a number between 0 (low) and 1 (high).

to start, we can make this a normalized percent based on the access logs or iRail

Problems I see

Bertware commented 9 years ago

I'm working on this, based on the amount of stopping trains

pietercolpaert commented 9 years ago

Awesome! Looks like a very good way to do it!

I'd call the column name vehicle_frequency: how much the station is frequented by a vehicle, such as a bus or a train.

What do you think?

Bertware commented 9 years ago

That's okay for me. I created a php script to calculate the frequency, going to look at the sorting tomorrow. Where should I put the frequency calculation? (it's a file that should only be run once on a .csv file to add the extra collumn)

pietercolpaert commented 9 years ago

If it's a PHP command line thing, let's put it in bin/frequency_calculator.php?

pietercolpaert commented 9 years ago

@Bertware any update on this? (I don't want to rush you though, just checking whether you're still on this)

Bertware commented 9 years ago

I'm waiting for the open data descriptor thingy needed for build.js, was going to ask you about it one of these days :)

pietercolpaert commented 9 years ago

Right! If that's keeping you back, then let's use a dummy for now: http://example.org/vehiclefrequency

Bertware commented 9 years ago

Allright, that should get me on track again!

pietercolpaert commented 9 years ago

@Bertware Fixed it quicker than I thought: use http://semweb.mmlab.be/ns/stoptimes#avgStopTimes :)