02 September 2011

Microsoft Academic Research: First impressions

Microsoft Academic Research is a new service that continues the company’s attempts to catch up to Google’s services – Google Scholar in this case. I hopped over and typed in my name. Not just out of vanity, mind you, but because I know what should be returned.

The search allows you to select certain domains. A lot of my research is spread over a wide field, so I checked as many boxes as I thought might be even tangentially relevant.

Oh dear.

I got five hits: Three in neuroscience and behaviour, one in biology and biochemistry, and one in clinical medicine. (Wha...?) Even for a beta version of the service, I was expecting double digits at least. Maybe I should have checked more boxes.

And... what a second... who’s that guy?

I have never met Michael N. Nitabach. He wasn’t an author on the paper. A little clicking reveals that there’s an “Edit” button, and that I can remove him as an author:

What else can I do here? Ah, there’s a spot to add a PDF link. Since this paper is open access, I can do that. Easily, let’s grab the DOI and head to the page...


Why is the DOI taking me to Figure 7 instead of the main page?

Okay, let's fix the DOI for the article. And add the link to the PDF. Might as well copy and paste the abstract too while I'm here. Why am I doing this work again? Isn’t the point of the database to have this stuff for you?


Also noticed that the “type” of article includes “poster”.

While I am a big poster booster, I don’t know that I want posters, which are typically very gray literature, in an academic research database. But Google Scholar catches blog posts sometimes.

Back to the main page. Hm. What are these “Were you looking for these authors” bar along the top? Yup, that’s what I was afraid of. Each one has pulled a different set of papers, even though all three of the alternates specify the same name. Why is the search for my name returning four hits instead of one? It’s not as though one was “Z Faulkes” - they all have the same name spelled out, although one includes my institution’s address below it.

The third “were you looking for...” – the one with my address underneath – is interesting, though.

This profile has a nice little dashboard, an RSS feed for updates... still missing a buttload of information, though. Let’s try editing author information. Ah, here is where add pictures, my home page, and - aha! - I can merge the three different profile!

But it’s going to take a week to verify the merger. While I appreciate that I can submit corrections, there is going to be too much literature to crowdsource corrections.

My first impression of this service is to ignore it for about a year, or until I hear about a major update. There’s just too many odd and unpredictable things going on here. It’s not trustworthy yet.

Hat tip to Bjorn Brembs on Twitter.

Additional: Bjorn Brembs followed up, saying:

It's less than beta and some/much of what you mentioned they are aware of.

No comments: