We're relaunching on Monday, and we'd love your thoughts!

[Jason Priem]

Hi All,

The Vancouver megasprint has been going pretty awesome so far. We're planning a huge relaunch on Monday, and with a new name: ImpactStory. We're also doing sort of a soft launch this weekend, so all y'all in the loop can get a SNEEK PEEK at the awesome new features live on the site right now.

We'd love it if you could take a look and let us know your thoughts this weekend. Try creating a report of your own if you get a chance...it's really easy now with importing from Google Scholar profiles. Or if you're short on time, check out the sample report.

Since the new site is live now, it's not really a secret; that said, the weekend is generally low traffic and we're sort of counting on that as we work out bugs. So don't publicize it widely till Monday if you don't mind. Then tell All The People :)

Thanks, and we look forward to your feedback!

j

--
Jason Priem
UNC Royster Scholar
School of Information and Library Science
University of North Carolina at Chapel Hill

[Chris Maloney]

Hi, this is really nice!

I have one usability niggle:  I created an impact report for myself (http://impactstory.it/collection/o8cndf), then went back to the home page, and couldn't figure out for a long time how to get back to my report.  I finally figured out that if I click on my username (email) then I'll get a list of collections, but it is not at all obvious.

Another issue I have is with the github report.  I don't know if anybody has brought this up before, and I don't know what the solution is, but it seems to be based exclusively on my own repositories.  What about other repositories that I collaborate on?  For example, I originally started the DtdAnalyzer under my own account (Klortho/DtdAnalyzer) but it has since moved to NCBITools/DtdAnalyzer.  I'm still the main contributor there, but it won't show up in my impact report, right?  I think that's a problem, because it's a slight perverse disincentive to collaborate.

Again, this looks really nice, great job so far!

Chris

[Jason Priem]

Thanks for the feedback, Chris! We'll think about better ways to let folks view their reports...maybe a "my reports" link on the navbar or something like that.

I'm not sure if GitHub exposes the kind of information you're talking about. If so, I would certainly be great to pull that in. We'll take a look.

j

[Martin Fenner]

This looks great, congratulations!

Something that confuses me is percentiles for very low numbers (e.g. 1 link to Wikipedia, 2 tweets). But the new grouping of sources and the percentiles are a big step forward.

Best,

Martin

[Jason Priem]

Thanks Martin!

I'm not totally sure I understand your question, but I think you're asking how something can have just, say, 1 wikipedia citation and still be in the 90th percentile or whatever?

If so, I agree: that's confusing. But it is (if we've done our job correctly) accurate, since we're counting relative and not absolute performance. 

Our sample of All Teh Articles (as randomly gathered from Web of Science) indicates that even a single Wikipedia citation puts you above ~90% of the articles out there. So then boom,  you're in the ~90th percentile, despite a low raw count.

This is, alas, a bit counterintuitive, but there's no getting around it as long we have metrics where most articles score a big fat zero. This may eventually improve for some metrics; I expect the percentage of tweeted articles to grow in the coming years, for example. In other cases, like F1000, the sources are always going to be selective and we're just going to have to get use to sparse data.

The worst part is that often a single event doesn't mean much--a tweet, for example. Even though it puts you in the 90th percentile or whatever, it could be from your mom (or more often, from you). In these cases, the content will be much more important than the percentile; if it's a doctor tweeting your study to her patients, that one tweet can support an important story to your funders. In other cases, a single event means a lot more--like F1000. Finally others, like Wikipedia will be in between.

I think an important challenge for ImpactStory will be to do a better job of finding and presenting the content behind the numbers, to better unearth the good stories behind low counts. We'll be putting a lot of work into that in the coming months. In the meantime, reporting the raw numbers along with the percentiles at least gives folks more context than they had before.

j

 a pretty good accomplishment, since fewer than 10% of articles attract one. That means you're in the 90th percentile (or thereabouts, depending on the year), surprising as it might be.

[Brian Hole]

I really like this - the new title and the breakdown of activity types is especially nice.

Please let me know as soon as the new version of the API is ready for use and I'll get our journal-side code updated. Can we assume that the older t-i API will keep working for a while as well?

Cheers,
Brian

--
Brian Hole
Ubiquity Press Ltd.

Gordon House
29 Gordon Square
London WC1H 0PP

www.ubiquitypress.com
www.twitter.com/ubiquitypress

Mobile: 0785 0769 510

[Paul Groth]

Hi,

It looks great. I'm not clear how I should add my google scholar info?
I just exported a bibtex and importanted it. Ist here some other way?

cheers
Paul

[Martin Fenner]

Jason,

yes, I'm talking about sources that only have data for a small percentage (e.g. 5-20%) of articles, so that for example one Wikipedia citation puts you in the top 10%. It would be good to indicate that somehow.

Best,

Martin