Slashdot Mirror
Top 10 Digital Cameras on Flickr
zokiii writes "This list represents the top camera makes and models used to create photos uploaded to Flickr. The list is generated automatically by periodically sampling the EXIF data from the stream of recent uploads." This is actually an incredibly simple idea, but a really useful one when considering a new camera to buy. Score three points for scrapers.
Beautiful use of statistics and just the sort of thing that the Flickr board of directors can take to their marketing meetings. Companies eat this stuff up as it is a real world reflection of specific markets using their products.
Interestingly, these statistics show that most of these cameras are solidly in the middle "pro-sumer" market and demonstrates that both Canon and Nikon did good when they went after a market that was primed to explode much like the 35mm market did in the late 70s and early 80s. Technology in both cases made it possible to build high quality cameras at affordable prices.
Of course this means good things for the consumer, but watch out for a new megapixel war much like what we saw with computers and the Mhz war. Of course just like the Mhz war, the MP war is going to be mostly hype as I've seen some damn fine images from 6 or 8 MP cameras that were far above the quality produced by some higher MP count cameras. The secret is going to be the size and quality of the individual imagers on the CCD, the quality of the image management chips (I tend to prefer Canon's DIGIC) and the quality of the lens.
Visit Jonesblog and say hello.
They mention that stats might be skewed, by the sheer volume of photos uploaded by photo enthusiasts. The implication is that they're simply totalling up all the pictures.
Wouldn't it be more accurate to generate stats by user? (ie: x% of flickr users have camera Y)
Well, err...
It tells you what other people are using. By this logic I should be driving a Ford Mondeo (and using Windows on my PC!)
init 11 - for when you need that edge.
WTF is a "digital camers"
Interesting idea, but there is a problem with the methodology. The app samples about 9,600 pictures per day. Since multiple photos from the same user are included, it will skew more heavily towards users who take & store (vs. take & delete) and those who are pro users (since they can upload more). This behaviour is more likely found in someone who fits the 'prosumer' profile, hence the abundance of these type of cameras on the list.
For version 2 of the stats, it would be really useful to have it ignore multiple camera models from the same user. It would also be neat to see it compare post-processing apps and general camera 'categories' (i.e. cell phone vs. p&s vs. slr).
When you have nothing left to burn you must set yourself on fire
How are there 27 comments already and nobody has mentioned a typo in the HEADLINE?
s/Camers/Cameras
I'm starting to lose faith in slashdot...
"The cup is in turn designed for holding hot or cold liquids, and has an open rim and closed base." --US Patent #5425497
It's just a list of which cameras produced the most pictures uploaded to flickr. The article itself points out that it's likely skewed because enthusiasts that spring for expensive cameras are also more likely to spring for premium flickr accounts where they can upload an unlimited number of photos.
So what does this list tell us?
The people who spend a lot of money on their 'prosumer' cameras take a lot of pictures. Well, they would have to, otherwise they just wasted that money on their prosumer camera. And they're also more likely to spring for a premium flickr account. Well, they would have to, what with all the crappy pictures they have to take to legitimize buying an expensive camera.
Even looking at the individual users pictures tells you nothing about the quality of their camera. They might just be really good photographers, or they just throw out 9 out of 10 shots because the camera sucks.
These statistics are pretty much useless to anyone. They're only useful to flickr itself; they can now dream up new ways of advertising, like show ads for camera accesories based on EXIF info.
Now, if they had a list of camera models by number of users, that would be more interesting (probably more accurate portrayal of marketshare than the manufacturers themselves give off). Or of models that are used to post the fewest pictures before the users photos suddenly change to another brand (in other words; quickly abandoned crappy models). Models that are used most for dark scenes with/without flash. That sort of thing.
SCO employee? Check out the bounty
To allow this totally BS use of statistics to actually MEAN something.
As one other photographer already mentioned part of.
1. Pro photographers are not going to store info on flicker
2. Pro photographers are going to be outnumbered by non pros, thus changing which is the best "camera"
3. It has been proven that a good photograper can take a better photo with a disposible camera than many consumers can with the "best" camera.
4. Others have noted that some cameras are lumped together, and others have multiple categories by the way the data is sampled.
5. No ratings of which photos are best are factored intot he sampling.
Sure, mod me down for this, but it doesn't change the fact that this story is totally BS.
The top 4 cameras are sub $1000 cameras. I realize everyone isn't prepared to pay $800 for a camera, but they aren't professional. I would but them in the high end consumer, amateur photographer category, so yeah I would say it's useful. I own a 350D myself. It's my first Digital SLR camera and I am very happy with it. A list like this would have influenced me.
Promote Charity on Myspace, Show Your Colours!
Yes, and web statistics show most people use Internet Explorer so that must be the best Browser.
Some drink at the fountain of knowledge. Others just gargle.
Really??? You'd base your choice of an $800+ camera on a list that doesn't give any indication of photo quality or camera usability, but simply indicates the mass volume of pictures taken?
No, it would be far better to do more conventional research. I would never consider such a useless statistic.