Comments on: Flickr photo uploads by camera model

By: jacreative

jacreative — Tue, 10 Oct 2006 15:39:34 +0000

I disagree with your assertion that someone adding 2000 photos in a given day would not affect the rankings. Regardless, it would be interesting to sample a larger size and break down the results by distinct username and camera model. This would get to some of the earlier commenters questions about popularity of cameras (as in # of photographers) versus number of photos taken. Maybe D50 users just take (or upload) more photos per user.

By: fervus

fervus — Wed, 27 Sep 2006 00:18:24 +0000

It might be interesting to be able to see most popular cameras for specific tags, or even be able to search for a tag and see which cameras had posted the most images with that tag, and so on?

Just a thought…

By: quadrapop

quadrapop — Thu, 21 Sep 2006 23:44:42 +0000

the problem with interestingness is that you get lots of photos of cas and flowers;-) followed closely by landscapes, so you will get a skewing towards photographers who are good at cats, flowers and landscapes…

great idea to sample the data this way, well done FD

By: Sabik

Sabik — Thu, 14 Sep 2006 04:08:00 +0000

Seconded on the request for weighting by popularity (presumably that would be by Flickr’s interestingness).

While it is true that “cameras don’t make photos—photographers do”, this would give an indication of the cameras used by good photographers. That’s probably a better measure of a camera than the raw number of photos…

By: rich

rich — Sun, 27 Aug 2006 22:23:49 +0000

Make no mistake this is an interesting number.

My point was that without disclosing what your sampling methodology is, what your controls for randomness, etc it is difficult to understand whether your analysis supports your conclusion. Stating conclusively that something ‘would not affect rankings at all’ reveals a certain naivete about statistics and quantatative analysis.

Probably not worth spliting hairs about. Thanks for the work to date.

cheers.

By: John Watson

John Watson — Sun, 27 Aug 2006 19:26:58 +0000

The method used to gather the data is to sample the recent uploads stream of all users. Flickr gets between 500,000 and 1,000,000 photos uploaded every day. The weekly ranking is for an entire week’s worth of data. One person uploading 2,000 pictures in a day (or even every day for a week) would not affect the rankings at all.

By: rich

rich — Sun, 27 Aug 2006 13:25:13 +0000

Do these numbers reflect the popularity of the camera or how frequently owners of respective cameras upload to Flickr? Seem that if one person uploaded 2000 pics in a day and other (different) camera owners uploaded a handful then the numbers might be skewed. Not sure from the description of the methodology if this is accounted for…

By: Fergus Macdonald

Fergus Macdonald — Fri, 25 Aug 2006 15:17:13 +0000

What a great idea…kudos FD. Glad to see that the number one spot goes to the D50. I often think that people look down on it because it’s got a lower number than the D70s…but it’s almost a better camera!

It’s cheaper, lighter, smaller and better looking. It has newer parts including a better motion tracking auto focus. However, it writes to SD card at roughly half the speed that the D70s does and lacks a few features and buttons…which can make it harder to change some options quickly without having to go into the menus.

If you are thinking about getting one…i definately would. Either the D50 or the new D80.

By: David Merwin

David Merwin — Thu, 24 Aug 2006 22:01:32 +0000

Photodoto » Blog Archive » Flickr photo uploads by camera model…

Photodoto » Blog Archive » Flickr photo uploads by camera model
Pretty cool graph showing the top cameras used on Flickr. I am thinking about buying a Nikon D50, so this is of particular interest to me.

……

By: John Watson

John Watson — Tue, 22 Aug 2006 16:54:48 +0000

But cameras don’t make photos—photographers do. :-)