2

I am interested in understanding which US airports are used by various airlines. One could visualize this with a spreadsheet:

enter image description here ​ ​

This examples shows that Southwest flies out of Midway, but not O'Hare. Delta flies out of O'Hare, but not Midway. Both fly out of Baltimore.

I'm wondering if there is a clever way to tease out a matrix from openflights.org.

A spreadsheet would be preferable although, I suspect that if the data is out there, it is stored in a RDBMS.

  • 1
    the 'teasing' is actually just programming, so I'm marking this to close because it's actually a programming question. More details below – philshem Jul 30 '16 at 7:12
  • i'm confused because gatorback is asking for a spreadsheet. i'm assuming it would take programming to make said spreadsheet? – albert Jul 31 '16 at 19:16
  • The spreadsheet is merely a visualization tool to present data examples. Other useful visualization \ analytical tools include MATLAB. That data is really the focus and the encoding of the data (CSV, .mat, etc). This is not a programming question of how to visualize the data. I am indifferent to how data is stored / encoded, though standard encoding simplifies loading data into spreadsheets and other analytical tools – gatorback Aug 1 '16 at 5:38
3

The data you are looking for is already at openflights.org/data. The reason it's not displayed as a matrix (pivot table, non-normalized, crosstab) is that there are 9'541 listed airports and 19'845 listed airlines and creating a sparse matrix is neither good for storage space, performance, user-queries, etc. One reason is that you've used a piece of data, the airport name, as a column name. So you can't easily make a dynamic filter.

Anyway, to create a matrix you'll need to know some programming, or a spreadsheet tool like Excel.

First, download the airlines.dat file. Each airline is given a 3-letter code (key).

Then, download the routes.dat file (2 MB) and use the airlines.dat 3-letter code to join the airline name to the airline code.

Then you need to de-normalize the data, which is a fancy way to make a pivot. Each SQL-flavor, excel, and programming languages can do this in their own way.


In my opinion, because the data is easily found and with a friendly license, I'm going to suggest closing the question. I'd recommend starting with a programming or Excel solution and then ask in the appropriate forum if you need support.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.