Saturday, March 29, 2008

Flickr Walkabout Intro

I'll start with an Introductory Mosaic from my own site, recently set up as part of the Admin team Feature for the blog of "Damn! I Wish I'd Taken That!!!"

I am often asked about 6th Lake, as I am one of the most-Googled chroniclers of this gem in the Central Adirondacks.
I make sure to tell everyone it's beautiful but the fishing sucks.

Speaking of Google......
By some quirk of fate, this image has risen to the first page, first row of the Google Image Search for "The Two Ways Of Life"--among more than 25 million links. I have had my personal print for 40 years, and was only prompted to scan it for my Flickr site.....

....when I ran into this image by Fabiola Medeiros during my first week on Flickr. I imagined them paired for observation...........
.....which led to a Flickr group created to display the two images together. Fabiola was pleased by the views and the comments that came her way. I kept the group going so the images would always have their home together, but changed it to one where people new to Flickr can find useful information about Flickr Groups: Understanding Groups

Speaking of Image Searches......
.....this image somehow rose in the Yahoo Image Search results for "Duck," (first page, 3rd row, among 2 million links) and gets clicked daily, making it the third highest in my image views.

Though there are many images where he plays a part that are much more fun......the first shot of the series remains the one that gets the attention.

Speaking of Yahoo Image Search, this image of Rockport, Ma. "Motif #1" has reached as high as the first page among 62,000 links, and currently resides on Page 8. It's difficult to believe that an image so recently posted to the internet can rise so quickly in the Search results.Motif #1
The arbitrary nature of the Search listings is part of this image's rise to Page 8 of the Yahoo Images Search results for "Rockport Motif #1" among 595 total images, while the first one ranks lower.
Boats in Harbor
This brings to mind that other great ratings game, Flickr's own Explore......

........which may be the single biggest mystery in the "world of photography." Here you have an online community with over 24 M accounts focusing at least a little attention if not outright fascination on the rotating series of 500 images rated "most interesting"--by a supposedly anonymous algorithm--and promoted whole-heartedly by Flickr Central.--so that it is nearly impossible to spend any time logged in and NOT know something about Explore. There are groups to discuss it, groups to boost the members chances of being Explored, groups to post the winning images, an independent website whose main claim to fame is making Explore banners for anyone who has had even 1 image selected. People display their banners on their personal sites and blogs........

ANOTHER ADDED, by wildcardpoet


they make SETS of their Explored images, and may actually spend time happily perusing the Explore pages hoping to find images to love.

An argument concerning Explore:

This image of mine was Explored the day after I set up my site. I knew nothing of image views and Contacts, Open and Invite-Only groups--in fact I did almost nothing else but upload images after I signed up--and had no idea that Explore had even happened. This is not a unique story, I might add.

So here's a question: how does an algorithm find an image and rank it and find it high in "interestingness" when there are no data from member interactions to be parsed? As I remember it, I stumbled on my first group invitations late enough into the second day that they almost certainly were sent to me because some people saw this one image deep in Explore.

Here's another question: how is it that porn is 100% screened out? If ratios of views to comments and faves--with various ratings of Flickr strength thrown in, plus other technical data given changing weights over time--are to account for the entire process of picking Explore images, wouldn't something offensive eventually get through? Just once?

And another thing: how is that certain people seem to have the mojo to get literally hundreds of selections, while the vast majority will never have one? How can their data always make it to the top ranks of the algorithm?

The simplest answer, following Occam's Razor, is the presence of human eyeballs somewhere in the Explore process, probably more than once. That's all I have to say.........

...........speaking of my first entry, it isn't a great image, but it's not horrible, either; that pretty much sums up several hundred of the 500 Explore images available 24/7/365. Still, there are fantastic images captured with flawless craft being displayed at any given time in Explore, so it is not to be dismissed as a site-sponsored deification of middle of the road achievement. To my mind, displaying these awe-inspiring images justifies the entire corporate endeavor. The more mysterious the process of bringing the images forward, the better off everyone is--I have to believe, since it's working so well.

This one was posted later on and garnered no interest at all from Explore or anyone else.

No comments: