City_distribution_in_HN_jobs - GitHub

Re:City_distribution_in_HN_jobs - GitHub by prolepunk on Tuesday August 02, 2011 @11:24AM (#2837432)

All the code does is looking up mentions of cities defined in the dictionary. How do you know that the cities aren't mentioned in any other context? How do you know that this is a resonable number of cities and their respected abreviations?

This is just lazy, and I'm not using word lazy as a technical term.

[ Parent | on HN ]

Re:City_distribution_in_HN_jobs - GitHub by codeherb on Tuesday August 02, 2011 @11:28AM (#2837450)

I think while all the failings you point are valid, the basic analysis he does is "good enough" for noticing trends. Good job!

Commenting and not actually fixing the code - now that is lazy :P

[ Parent | on HN ]

Re:City_distribution_in_HN_jobs - GitHub by prolepunk on Tuesday August 02, 2011 @01:17PM (#2837874)

Idea is good by the tool is not good enough -- it only notices the population centres that author has took a note of, this looks like unintentional confirmation bias.

See my reply to author with fixes.

[ Parent | on HN ]

Re:City_distribution_in_HN_jobs - GitHub by pama on Tuesday August 02, 2011 @04:22PM (#2838556)

I've added London, as per your transient request, which is now deleted. I also changed the comments to "select cities", as this code by no means represents a comprehensive analysis over all possible cities.

[ Parent | on HN ]

Re:City_distribution_in_HN_jobs - GitHub by prolepunk on Tuesday August 02, 2011 @05:22PM (#2838754)

I just pasted my original code with properly indenting in here, so I thought to remove ugliness once I copied it pastebin.

Also I saw some small city in california 'Los something...' with population of 28k which didn't get caught by the first function that only lists 50 most populous cities so I wrote the second one that includes every town in US.

[ Parent | on HN ]

Re:City_distribution_in_HN_jobs - GitHub by pama on Tuesday August 02, 2011 @11:53AM (#2837562)

This was indeed a quick hack. Because the who-is-hiring thread is focused, I expected a small false positive error rate. Let me know what additional cities or abbreviations you'd like to see (or fork me on github).

[ Parent | on HN ]

Re:City_distribution_in_HN_jobs - GitHub by prolepunk on Tuesday August 02, 2011 @01:15PM (#2837866)

ok formatting went all wrong, here's the same code on pastebin: http://pastebin.com/S5324Gfw

[ Parent | on HN ]

Re:City_distribution_in_HN_jobs - GitHub by pama on Tuesday August 02, 2011 @04:25PM (#2838567)

Thanks for the pastebin code. I've added London to my list for now, and I'd be interested to seeing the results with all possible cities if you wanted to merge your code in github.

[ Parent | on HN ]