Wednesday, June 11, 2008

Pubmed Faceoff: The Extreme MakeOver MashUp For Search Results

Friends/

A Most Impressive Development !

/Gerry

Pubmed Faceoff

What's this all about?

This site applies a simple, photorealistic variant of the Chernoff Faces visualization technique

[http://en.wikipedia.org/wiki/Chernoff_faces]

[http://kspark.kaist.ac.kr/Human%20Engineering.files/Chernoff/Chernoff%20Faces.htm]

to impact factor data for papers in the PubMed database of biomedical literature. Basically it allows you to search PubMed and have the results represented as a set of human faces. "I was published a couple of years ago in a crap journal and nobody is citing me."

The theory is that mapping multidimensional data (in this case the age, journal impact factor and citation count associated with each paper) to facial features takes advantage of the fact that our brains are highly tuned to recognise, process and differentiate between human faces.

"I was published a couple of years ago in a mid-tier journal. I've got slightly more citations than expected."

Each paper is represented as a face. The ethnicity and gender of the face is selected at random for visual interest - you can turn this feature off if you so choose.

"I was published recently in a good journal. Citations are as expected."

The age of a face correlates with the publication date of the paper. Younger faces are more recent papers. "I was published recently in a good journal and I'm getting lots of citations."

A smile means that the paper has been cited more times than expected (based on its age).

Larger smiles mean more citations.

A frown means that the paper has been cited far less than you might expect.

The raised eyebrows correlate with the impact factor (sort of - actually the Eigenfactor) of the journal in which the paper was published.

[http://www.eigenfactor.org/]

Source
[http://www.postgenomic.com/faces/index.php]

Thanks To

/ Frank Norman /Librarian / National Institute for Medical Research / The Ridgeway, Mill Hill, London NW7 1AA, UK /

10 comments:

Darlene said...

Really interesting stuff being developed. Wow.

Anonymous said...

too many faces
too crowded

Anonymous said...

I agree!! There are too many expressions to figure out which face goes with what citation factor. Numbers are easier to figure out. The person who came up with this idea was thinking too much!!

Anonymous said...

Yeah, me too, a peculiar feeling. Too many faces on my private desktop, a silent crowd waiting my response, I have nothing to do with them...

Anonymous said...

I did a test search. When the results used random gender/ethnicity, it was extremely difficult to discern the subtleties, that is, takes more time than it's worth, to evaluate what each face meant, in terms of age and expression. When I controlled for gender/ethnicity, the subtleties of the facial expressions/ages were more obvious, but still required what I thought was too much time to decipher what each face meant. Perhaps with practice, you get really good at reading these "faces" and interpreting their meanings.

Anonymous said...

How come everybody's bald?

Anonymous said...

Reminds me of the old "Max Headroom" TV show. Kinda creepy.

Anonymous said...

This is interesting, however, what about implicit racial, ethnic and/or gender biases that people have? For example, will people "not take seriously" a female or black face that indicates high impact? Think that we are all above that? See the Implicit Association Test which is part of the ongoing "Project Implicit" at Harvard and test yourself:

http://www.understandingprejudice.org/iat/index2.htm

https://implicit.harvard.edu/implicit/

I think presenting results in this fashion might be adding (or implying) a hidden layer of information that is totally irrelevant depending on what gender, age and/or race face that appears....

Anonymous said...

When I'm searching for citation information, I don't want to have to wade through 'kiddie' stuff but am an adult capable of reading difficult information like numbers. What's next? Emoticons and smiley faces? This kind of stuff may be better left to information targetted at the k-6 level.

Anonymous said...

I thought this was a really easy way of seeing the impact of a journal, but agree that it only really worked when the face was the same gender/enthnicity (it didn't really matter which one, but the randomness made it too confusing).