1/2 million photos analyzed; what makes a good photo?

[Derrel]

I just saw this story linked to via The Online Photographer, and thought it was kind of interesting to see the results OkCupid, and online social network and dating site, found after it analyzed 522,000 photos.

Don’t Be Ugly By Accident! OkTrends

Their article showed that 4/3 cameras made the most attractive photos. They found that using a flash adds an average of seven YEARS to the apparent, perceived age of its subject! Interchangeable lens camera photos lead to more positive ratings of photos than those done with point & shoots, and P&S photos were better than cell camera photos.

In a possibly related vein, it seems that wide-aperture photos, with a sharp person and a blurred backdrop, get the most positive ratings, and wide apertures like f/1.2 to f/2.8 all score well into the "plus" or "positive" side, while small-aperture photos like those shot at f/11 to f/22 fare poorly by comparison.

The article has some other interesting bits. I have no idea about the validity of the statistical analysis done by these folks, but I found their conclusions interesting to look through, since there are so, so few examples of photos and quantitative analysis available for discussion.

 

[supraman215]

Surprisingly few examples considering nearly all digital cameras offer some sort of exif at least for shutter and aperture. It's wonderful now how all these values can be quantified even most cheap P&S cameras provide this info in the photo. The data analyst in me gets excited.

I notice how you didn't mention the first graph in your link. Derrel I think you're going to insite a major Nikon, Canon (Leica) battle here. Maybe that was your secret intention .

Maybe there is some truth to the "hot chicks with canons" as posted in a previous thread. I'm a little surprised at the lack of any break out by gender. One would think that would play a role since the gender of the person determining attractiveness is relevant. I think the n of the camera brand would be relevant as well.

 

[Derrel]

Surprisingly few examples considering nearly all digital cameras offer some sort of exif at least for shutter and aperture. It's wonderful now how all these values can be quantified even most cheap P&S cameras provide this info in the photo. The data analyst in me gets excited.

I notice how you didn't mention the first graph in your link. Derrel I think you're going to insite a major Nikon, Canon (Leica) battle here. Maybe that was your secret intention .

Maybe there is some truth to the "hot chicks with canons" as posted in a previous thread. I'm a little surprised at the lack of any break out by gender. One would think that would play a role since the gender of the person determining attractiveness is relevant. I think the n of the camera brand would be relevant as well.

No, no secret agenda...it's fascinating to see the metrics they selected. I thought it was interesting that iPhone users get more sex than users of other phones!!! ANybody care to address that one???

Hot chicks with Canons...I musta' missed that one...

 

[shaunly]

So too look your BEST, shoot with a Panasonic 4/3, Ultra shallow DOF, NO flash, at around 6:30AM. ohh, and use an Iphone.

 

[Mike_E]

Interesting, but totally specious.

The math used in this is so irrelevant that I really doubt the claimed number of photos is even accurate.

 

[Scatterbrained]

One key question is where did the get the photos for this? If they are just skimming facebook then this test is already irrelevant. Of course the accountant in me cringed when they said "we had data on all of these people so we just dumped it in excel...." voila!, instant concrete science. I guess I wasted a lot of time on my degree, all I really need to do is pump numbers into excel and see what comes out. :er:

I will admit though, it is a humorous study.

 

[Derrel]

One key question is where did the get the photos for this? If they are just skimming facebook then this test is already irrelevant. Of course the accountant in me cringed when they said "we had data on all of these people so we just dumped it in excel...." voila!, instant concrete science. I guess I wasted a lot of time on my degree, all I really need to do is pump numbers into excel and see what comes out. :er:

I will admit though, it is a humorous study.

They mention where the information came from: their own web site and their own users. As they stated: " As always, our data comes from dating site OkCupid, one of the largest, and most interesting, datasets on the web. This article aggregates 11.4 million opinions on what makes a great photo."

Funny that they consider their own dataset one of the most interesting! This article is from the site's own ,official blog. The site claims that it has 3.5 million users, and they regularly analyze their data and blog about it.

 

[supraman215]

One key question is where did the get the photos for this? If they are just skimming facebook then this test is already irrelevant. Of course the accountant in me cringed when they said "we had data on all of these people so we just dumped it in excel...." voila!, instant concrete science. I guess I wasted a lot of time on my degree, all I really need to do is pump numbers into excel and see what comes out. :er:

I will admit though, it is a humorous study.

Science? Maybe market science but hardly. I'd call it market research.

 

[David Dvir]

Any photo that makes you stop and think .

 

[Garbz]

Saw this a few weeks ago and it really made me laugh.

But what's more likely, Panasonic cameras take better pictures, or people who buy a Panasonic instead of an ultra expensive DSLR are likely to be less nerdy and thus draw in hotter friends

 

[Derrel]

Saw this a few weeks ago and it really made me laugh.

But what's more likely, Panasonic cameras take better pictures, or people who buy a Panasonic instead of an ultra expensive DSLR are likely to be less nerdy and thus draw in hotter friends

So, what you're sayin' is Panasonic owners are sexier!!:thumbup:

I'll go along with that!