OpenSocial Gadget Contest

Adding new features to Profiles and/or VIVO

Printable Proposal Content with Comments

Use your browser's print function to output proposal content only, with each proposal starting a new page. Print to Adobe PDF to produce a file. (Note: Chrome and IE9 do not support starting a new page for each proposal.)

Identifying resourses as we type.

Proposal Status: 

I think we can do a good job at identification of resources as we feed new content on VIVO profile. I had this idea last year. For instance, a researcher would enter his description in VIVO and as he types, he would get suggestions on VIVO resources identified in his statements.

 

I developed a prototype last year : http://entityfinder.appspot.com/

 

I could have bugs in there but I guess it's good enough to convey the idea.

Comments

Do you have ideas on where to integrate entity extraction functionality into a researcher's Profiles/VIVO workflow?

As a first pass you could make an AJAX call as the user types in the description box and fire a SPARQL that does the autocomplete for resource labels. Later you could try to improve and tie the javascript app with Opencalais to identify entities by grammatical parsing of the sentences ( after the user hits submit button as shown in the example link of idea description). In both the cases you could restrict the search to VIVO resources. Does that help ? I can provide you with the SPARQL autocomplete query to start with.

As a result of technical feasibility review conducted by the Virtual Home team at CTSI, this proposal was considered beyond the technical scope for implementing it as an OpenSocial gadget. On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Harmonize Profiles Gadget

Proposal Status: 

A researcher may have a primary, home location, e.g., UCSF Profiles, where they've taken care to make sure their profile is accurate and complete.  The researcher may also have a secondary profile at another institution, at an independent research lab, at a hospital.  Duplicating and keeping an extensive profile containing many publications, grants, patents, etc. would be tedious -- in fact, so tedious that few people would devote the effort.

 

What if there were a gadget that could be dropped onto the user's profile page that examined a primary and secondary profile and noted the differences (for a visual, think WinDiff or a source code merge tool).  When clicked, the gadget could do a merge or "harmonize" the two profiles, note the desired changes, and allow the user to approve.  For an empty secondary profile this would essentially be a copy operation, with the exception of links or reference types that may not be supported in common among two systems.  Afterwards, this Harmonize Gadget would be used to keep profiles aligned with the fewest number of clicks.

 

An initial test case could be developed between Profiles, VIVO, and KNODE. KNODE discovers researcher content differently than either VIVO or Profiles.  A researcher could claim their KNODE profile, then use the Harmonize Gadget to merge and align the machine generated profile in KNODE with the curated version in Profiles or VIVO.  Or a claimed KNODE profile could be used via this gadget to populate a VIVO profile.

 

This would be based on VIVO RDF retrieved from both systems via the gadget.  RDF would make the comparison of content much easier.  The merge and update back to each system is probably the biggest, new technical challenge. 

Comments

It would also be interesting do discuss a similar merge / sync profiles capability with ORCID once it is launched.

What are the pros/cons of user-initiated synchronization, vs. a client-server model, e.g. having Knode automatically pull down RDF from Profiles or VIVO, and occasionally poll for changes? Did you imagine synchronization as being a largely one-time process, or something that users do on a regular basis?

ditto similar comment on ORCID gadget proposal:

 

This sort of profile "sharing" is just getting going.  So in the beginning, what I'd think to have is a uni-directional ability to upload publication lists, etc. from Profiles or VIVO up into KNODE to populate a profile claimed in that system.  Or visa versa to take content listings from KNODE and download into VIVO or Profiles.  Initially I'd think this would be manually initiated by the user.

 

As this all matures, you can imagine a more automated process -- perhaps email notifications of profile changes from one system or the other that need to be sync'd.  Especially as there are machine generate content updates.

As a result of technical feasibility review conducted by the Virtual Home team at CTSI, this proposal was considered beyond the technical scope for implementing it as an OpenSocial gadget. On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Conference Recommendation

Proposal Status: 

1. Harvest "Research Raven" into structured form (this may require getting permission):

http://www.researchraven.com/about-us.aspx

Microsoft Academic Research has this data but it seems limited to computer science.  PubsHub also mantains such data but it's proprietary.

 

2. Using keywords from a user's profile, match people to upcoming calls for papers at conferences. Sort into two categories: conferences to attend and conferences to submit to.

 

3. List the top ranked upcoming conferences for a given person.

Comments

Paul - interesting idea. Given that most profiles have a very large number of keywords, how would you propose choosing which keywords that would be searched upon for a particular researcher? How would the 'top ranked upcoming cnoferences' be defined? I'm also curious if you have a sense for how much time, in general, that researchers spend trying to identify conferences to attend and submit to? If deemed valuable by reearchers, this might be a list that could be pushed to the profilees rather than them going to to get the list on their profiles.  Just a thought..

Here is one workflow:

1. Match keywords to ResearchRaven's categories. ResearchRaven lists 364 categories. That's 364 strings to match against the titles of publications, grants, clinical trials, etc. You could also match against more verbose fields such as abstracts. Here at Weill Cornell, our CTSC manually assigns one of 200+ areas of expertise from the National Center for Research Resources to each of our researchers. So if you have something like that around, you could match on that basis.

2. Sort conferences of a certain date range (e.g., 3-5 months from now) by number of times the categories it has been assigned matches the keyword corpus of your researcher. You could give more or less weighting depending on recency, whether it was in the title vs. abstract, etc.

3. Present a feed of conferences from Research Raven in that category.

4. Give users the opportunity to mark as "relevant/more like this" or "not relevant/less like this" for each conference. In those cases, increment or decrement the point value of that category accordingly. You could also upweight conference suggestions in the same category that have been marked as relevant by other researchers.

To answer your other question, I think conference recommendations are in the same boat as researcher profiles. They tend to get the most attention from researchers who are new to their profession.

What you say about recommendations being pushed to researchers rather than showing up in a profile makes much more sense.

 

Thanks for the example workflow Paul. Having a option for researchers to rate relevancy is an interesting one but not sure how likely it is that they would engage. Overall though, it seems that your idea could be a useful resource - especially for junior investigators, as you mention. Good luck!

...

Re the recommendations, another option would be to put them on the profiles with default or option to push

 

Hi, everyone. This is a fascinating discussion. I am the Web adminsitrator of ResearchRaven, so you can pretty well assume I am delighted to read all of your comments!

I am a bit confused (and I appreciate your patience with me here as I am not a scientist or an information scientist) by Paul Albert's 's comment here, "Present a feed of conferences from Research Raven in that category." Given that each category in ResearchRaven can be subscribed to via RSS or subscribed to via email alert, I am not following what would be new--except that I do see the value in rendering the info in ResearchRaven even more useful via the idea Paul suggests here, "Give users the opportunity to mark as "relevant/more like this" or "not relevant/less like this" for each conference. In those cases, increment or decrement the point value of that category accordingly. You could also upweight conference suggestions in the same category that have been marked as relevant by other researchers." I think Rachael Sak has a point when she says it might be difficult to get reseachers engaged in "relevant/more like this" or "not relevant/less like this" activity though some might if they are especially eager to follow some particularly specialized subject. And there might be some funny (or deplorable) examples of vendettas and gaming of the system.

It is hard to know what say about "recommendations being pushed to researchers" given in the info overload many of them speak despairingly of.

In any case, the RSS feeds are freely available for every category on ResearchRaven, so I hope everyone feels free to experiment with those if doing so would be useful.

Anyway, thank you for providing this stimulating forum. I am quite proud of ResearchRaven but must confess that none of your marvelous ideas had ever occured to me. Dang!

By the way, I do want to mention our sister site ScanGrants, a free online listing of grants, fellowships, prizes, and scholalrships in the health sciences http://www.scangrants.com/

Hope, thanks for your work with ScanGrants and ResearchRaven. Opening up that data allows outsiders like us to explore interesting mashup ideas, without needing to get pay or get pre-authorization — enabling innovation at the edge of the network.

Hi,  Anirvan. Well, I must say, having read your profile, that you have a pretty impressive track record when it comes to innovation. And interesting mashup ideas sound great to me!

See below

Selected comments from Reviewer(s): "Interesting idea, needs more development. Conference attendance has probably MORE to do with social and professional networking than scientific content (although this certainly needs to be there too)."

 

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Citation alert using data from Microsoft Academic Research

Proposal Status: 

Give profiled researchers a list, in order of recency, of citations to their publications as listed in their profile. Use Microsoft Academic Research's API to do so.

At present, the only way to know this information is to sign up using Scopus or Web of Knowledge. Google Scholar also has a tool but their author disambiguation is relatively weak.

If indeed you can get the names of the publications from Microsoft Academic, this has upsides for both system administrators and profiled researchers. Profiled researchers learn about who is discussing their work. System administrators, and the university in general, have a carrot to encourage researchers to keep their publications accurate and up to date.

Comments

Cool idea. Have you used or do you know much about the Microsoft Academic APIs? Work well? Stable? Just wondering.. we haven't ever used them.

I haven't used this API. This suggestion was based on a conversation I had with the director of scholarly communication at Microsoft Research. As you see, I sent an email requesting access to the API, and cc'd you on it. Let's see how this works out.

This is an excellent idea, if we can get API access. I wonder if the data might be better presented inline within the list of publications, like this:

  • Title of paper. (2 citations)

…rather than as a separate gadget.

Selected comments from Reviewer(s):

Reviewer 1: "Excellent way to increase the value of your research through collaboration with others working in complementary areas."
Reviewer 2: "investigators already get this service (weekly or periodic reminders of people citing your work).  Don't think it's much of a game changer.  Don't see the innovation."

 

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Organizational name lookup using Virtual International Authority File (VIAF) service

Proposal Status: 

About:

The Virtual International Authority File (VIAF) is an international service designed to provide convenient access to the world's major name authority files. Its creators envision the VIAF as a building block for the Semantic Web to enable switching of the displayed form of names for persons to the preferred language and script of the Web user. VIAF began as a joint project with the Library of Congress (LC), the Deutsche Nationalbibliothek (DNB), the Bibliothèque nationale de France (BNF) and OCLC.

 

Read more:

– Demo of the service: http://viaf.org/

– API documentation: http://www.oclc.org/developer/documentation/virtual-international-author...

 

 

 

Comments

Can you describe how the transliterated names might be presented on Vivo or Profiles pages?

At present, when it comes to UMLS and GEMET, VIVO requires users to click to look up by controlled vocabulary. My suggestion here is that when there is occasion to enter a name of some kind - especially if it's an organization - that Profiles present a dropdown menu of existing options from Profiles itself, and below that, those from VIAF. Or, you could only show VIAF options when Profiles has no such object. Or you could have a button that says "Lookup organization."

We can't get users to manually enter controlled terms perfectly. But, this way, we at least increase the odds that a user enters the same organizational name many others are using.

Great idea. This might be something that the VIVO/Profiles teams might want to take up, and bake deeply into the software.

As a result of technical feasibility review conducted by the Virtual Home team at CTSI, this proposal was considered beyond the technical scope for implementing it as an OpenSocial gadget. On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Rewarding a Profilee's External Media Presence/Q Score by Adding "Mention Mapper" or Profilee Rating

Proposal Status: 

One way to further engage the profilees and encourage their media (online and traditional) interaction: Might we add links to media "mention mappers", which quantify social (and other) media appearances, similar to how Q Scores rate familiarity levels of brands, companies and individuals? A good prototype is the Klout influence scoring model. In this way, we could reward profilees for their high "influence" scores in external, non-UCSF media. We could also consider, in developing each profilee's rating, whether or not s/he has added his/her UCSF Profile to his/her LinkedIn or Facebook profile.

Examples: SocialMention.com, Klout.com, BeEvolve.com, NCRR site listing media mentions

In this way, we could imitate those user-oriented and crowdsourcing "help" forums like MS Developers' Network/MSDN, etc.). These sites reward members for frequency (or helpfulness) of responses by providing "points", "stars" or other indices. It would also advertise, to profilees, the benefits of profile maintenance and frequent external media appearances. This would be a minimally labor-intensive way to transfer to profilees partial ownership of the task of promoting their own profiles. By encouraging a larger media presence, we would also broaden outreach for the UCSF Profiles in general.

Comments

There are several interesting ideas listed here... how about putting them in as separate proposals? that will facilitate discussion on each one, not to mention increase your chances of winning.

Done - I've reposted what had been the additional ideas from this posting, into separate proposals. Thank you very much for your suggestion.

It would be interesting to have a discussion on this topic.  Although KNODE is not looking to invent anything like a numeric Klout score, we are working on a innovative mechanism for expert relevancy ranking.  When you search for a disease area or biomed term, in addition to traditional search relevancy, other factors like journal prestige, author positions, cross references, etc., etc. would be considered.

What a wonderful idea, an expert relevancy ranking. Those parameters sound great, especially the author position (first, second, etc.). Great idea!

Interesting idea. Are there any open APIs that provide these kind of research-related metrics at this time, in bulk and without end-user involvement? I like Impact Story, for example, but they require users to register one at a time.

You might also want to take your idea of tracking media mentions, and split that out into a media mentions gadget idea.

Thanks for the suggestion, Anirvan. I'm sorry that I wasn't able to pursue that until today, after the closure of the improvement phase.

Selected comments from Reviewer(s):

Reviewer 1: "I think this could be a big hit if done properly--a mechanism for quantifying and sharing internal and external media hits, blog posts, letters to the editor, etc.  There is a WHOLE bunch of these types of activities that do not get measured or acknowledged by promotions committees or department chairs, but which could be illuminating and rewarding.  I think to tag the "mention" or "post" with a "reach" factor of the number of people likely to see the item would also be pretty cool.  This is all advertising/marketing speak..."
Reviewer 2: "the idea is good but this may be a very complicated project in terms of vetting what's appropriate to add publicly..."

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Cross platform research query

Proposal Status: 

Assumptions:
·Research scientists want to know what others have done or are doing in order to strengthen their knowledge base.
·Research information resides in various publications/data sets hosted in a variety of public or private domains.
 
Background:
A research scientist can use his/her own research data to correlate findings from other research scientists in order to determine who has worked or is working in this research area, with a focus on what others have studied, who performed the research, whether the researchers’ area of interest is relevant to their own research. If and when they find a match, they can look at the data set for further analysis if available. This type of correlation can also be performed within their own data set
 
Proposal:
Develop a way to enable a research scientist to query online publications and data sets based on their interest (e.g., key words, etc.). Allow the end user to define which platform(s) or other filtering elements as part of advanced search. Once queried, show the result in drill down fashion, starting name, title, contact information, etc. Further drill down will show publications and data sets. The level of detail is based on data set. Publically provided data from a platform like Pubmed will be readily available. But if the collaboration is established among universities and private companies then those could be used as well. This is really a matter of how much we collaborate within and outside.
 
Example:
user search on keyword: brca
 
The system searches on various platforms such as UCSF research DB, UC research DB, Pubmed, cochrane, medlineplus, Europe PubMed Central, Amgen, Genzyme, Genentech, Novartish.....
 
 
Level 1 result: Publication titles, names, contact information, etc.
 
Level 2 result: The whole publication
 
Level 3 result: Data set used
Level 4 Operation/Result: Do correlated query among others data sets with option of including your own

This enables the researcher scientist to contact the associated researcher. Then if allowed and available more reference data can be queried whether from system or directly from the source (the contacted researcher).

A tool like this could enhance medical research and collaboration tremendously.

The system could be as simple as providing names and contact information at first, and then expanded by adding more features to enable complex queries.

Just a thought!

Comments

I like this idea. It could be expanded to, say, specific lines of transgenic animals, etc. Given that publications can have DOI unique identifiers, why not have identifiers created for any dataset that gets used or referenced in more than 1 publication.

Yes! Datacite is an organization devoted to this purpose. DOI issuance must be initiated by the individual who provides access to the dataset, and just FYI, UCSF provides this service through a new website called DataShare in which datasets can be deposited in order to provide access, and these datasets are automatically issued a DOI and QR code.

This is an ambitious and exciting idea, but it might work best as a standalone project, rather than a plugin for fundamentally people-oriented systems like Profiles and Vivo. Are there existing systems like this that could be embedded or built upon?

Hi Anirvan, Yes, it might work best as a separate project but since we are looking at research based collaboration it might as well be a good starting platform for researcher to realise the possibilities of how far/deep the integration/collaboration can go.  I worked setting up Nextbio's [ http://www.nextbio.com/b/nextbioCorp.nb] building blocks which did similar things. They might have api to integrate with others by now. There might be other players on the market as well.

 

Yes, this is a great idea! However, if the proposed gadget queries only a set number of databases (as exemplified above), the additional resources that may be avilable elsewhere (e.g. in numerous other databases and repositories througout the web) would not be presented. So, a truly comprehensive result would necessitate a full web search; but maybe it is acceptable to limit results to only those found on some number of standard databases...

Also, Level 4 (performing an operation on the result) is again a fantastic prospect, but requires a level of integration beyong the "gadget" scale.

Sorry, didn't mean to be "Anonymous"

As a result of technical feasibility review conducted by the Virtual Home team at CTSI, this proposal was considered beyond the technical scope for implementing it as an OpenSocial gadget. On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

CTSI Data Autofeeds - Great Resources for Other UCSF Sites

Proposal Status: 

The automated CTSI data feed (and the UCSF Profiles mini-search feed) are very useful but not necessarily well-known features. The first feature auto-feeds live data from Pubmed publication lists and other UCSF resources into pages hosted on departmental or other UCSF websites. The second links from UCSF pages into UCSF Profiles. 

If we haven't already, might we ask UCSF IT leadership to broadcast these free services to a website managers' listserv or other target group? A "top-down" message from leadership to the IT community might help encourage broader use of this free service, with no financial outlay. Katja Reuters has already been very effective and proactive about promoting this feature:

From John Maa's Gen'l Surgery Dept. page: “View All Publications at UCSF Profiles (Powered by CTSI)” - the embedded link towards the middle of the page is a great example of use of, and credit to, the data feed

From UCSF’s Center for AIDS Research (CFAR)

From UCSF’s San Francisco General Hospital

This would add visibility both to profilees and to the CTSI's great resources for investigators. This visibility among both parties would hopefully facilitate further collaboration across the organization.

Comments

This isn't specifically a gadget concept, but I think it's a great idea.

As a result of technical feasibility review conducted by the Virtual Home team at CTSI, this proposal was considered beyond the technical scope for implementing it as an OpenSocial gadget. On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Adding "Mentoring Outreach" or "Collaboration Outreach" Star Rating

Proposal Status: 

Some sites like Yahoo Answers, the Microsoft Developers' NetworkAmazon.com product reviews and Yelp reviews, rate members based on frequency or helpfulness of their suggestions or reviews.

Is there a way to track how many times a site viewer emails a given profilee, after having searched using the pulldown menus? For example, after having searched for "Faculty Mentoring" in the pull down menu, a site viewer might use the contact info on the Profiles site to email a potential mentor. Ideally, the Profiles site would track frequency of emails (via Profiles) to a particular mentor, generated from this type of search.

This profilee would have a "Mentor Outreach Rating" which would show, say, three maximum stars for three or more emails. This would be similar to those crowd sourcing and review community sites which rate users based on, and thus reward members for, frequency of postings. This would encourage profilees to update their narratives and other text fields (on which "Faculty Mentoring" seems to base its search) to include info on "mentoring". It would also add visibility to the mentoring services offered by profilees. 

Similarly, if the site tracks the number of emails to a profilee, generated by (after) a search for "Global Health" (by specific country), we could roughly equate the number of email transmissions to a number of potential collaborations resulting from this search. The site would then generate a rating (three stars/three or more emails) for "Collaboration Outreach".

In each case, we would make it clear that the rating is based only on number of emails generated. Since many UCSF investigators are so metrics-driven, having a transparent measure of outreach efforts might encourage profilees to maintain and update their Profiles frequently. This would be a free and very user-friendly way to increase the visibility of, and traffic to the Profiles, while simultaneously rewarding profilees for frequent site visits and updates.

Comments

It wouldn't be technically challenging to track clicks on emails, and display them as a gadget. The faculty mentoring filter is a UCSF-specific addon, and not widely used at this time, so I'm worried that the numbers would be very low.

Selected comments from Reviewer(s): "Nice idea, but given implementation details may be hard …"

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Expand links (URLs) from citations to open access and data sharing repositories

Proposal Status: 

The current Profiles system only allows adding links (URLs) from citations entered manually, not through the automated PubMed feature.  Allowing faculty to add multiple links to each publication would be useful.  It would be even more useful to add tools to add links to open access and secure data sharing repositories.  In particular, these should link to open access repositories and supplementary materials or resources.  UCSF and other institutions have open access publication policies in which faculty are now required to give publishers limited licenses to their work instead of unlimited copyright. Data sharing requirements also exist for federally funded studies; links to secure data repositories or documents describing how to access the data would be helpful.  Some projects develop software programs, questionnaire instruments, training manuals, or other important research resources which journals may not publish due to space considerations or may not provide information or links to more details.

 

In addition, the PubMed citation list in Profiles should pull and display the PubMedCentral identification numbers (PMCIDs) that are assigned when the publication is deposited into the National Institutes of Health Manuscript System (NIHMS) open access repository, as required by federal law for federally supported research.   Having this information in Profiles would be very helpful for preparing biosketches for NIH Progress Reports and new grant applications.

Comments

This is a good idea, and might be best implemented as a core Profiles feature.

Agreed. These features seem to be integral pieces of system development, not so much a "gadget". However, the gadget could be something like an altmetric display for the researcher's productivity (e.g. displaying counts of PMID journal articles, DOIs for datasets and other pieces of 'gray literature') 

yes, that's a great addtion to my initial suggestions of enhanced functionality.  adding something like Web of Science's h-index would be very helpful too. building off your idea about gray literature, adding some way to count or link to things like policy papers that cite one's literature would be an excellent way to document T2 / T3 dissemination and uptake.

As a result of technical feasibility review conducted by the Virtual Home team at CTSI, this proposal was considered beyond the technical scope for implementing it as an OpenSocial gadget. On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

"deep text" searchable "Narrative" and "Faculty Mentoring" fields

Proposal Status: 

I have been promoting and using this tool predominately as a mechanism to connect mentees with career mentors within our home department.  Of course those who are interested in serving as career mentors must create specific text in both their "Narrative" and "Faculty Mentoring" fields of their profile.  These paragraphs provide a rich source of information about mentors beyond whatever words appear in their publication titles, keywords, etc.  For example, mentors can elect to disclose things about their personal life (e.g. sexual orientation, whether they have kids, hobbies, work-life balance issues, etc.). 

I would love it if the "search by research topic function" could "deep-text" search from these two fields.  Or perhaps there could be a separate search function if more feasible/viable?

 

It takes a team mentorship approach, so I believe this added functionality would increase the utility of the Profiles tool for mentees seeking adequate mentorship, especially to augment their research mentorship!  

Comments

Hi - thanks for the idea! The good news is that there is some functionality in UCSF Profiles today that might help you immediately, and perhaps can be expanded upon with your idea. Today, in a little known/used feature, there is a "Google-based" full text search of "Narrative" text if you search on by research topic. Here's how to find it: 1. If you submit your search, scroll down to below the search results table of people. 2. There is likely a box below the table of people that says "Full Text Search Results (beta) - Powered by Google" with a small + sign on the right. 3. Click on the + sign to expand the box and see results of a full text search that will (at this time) include search result matches from Narrative text and Titles of articles. If you search for a term that is not a keyword assigned to any articles (for example: "mhealth") -- there may be no people returned, but there will be results in the Full Text Search Results from Narratives. We were testing this feature out for utility, and would love to hear more back from you if you have more comments such as adding the faculty mentoring narrative (and perhaps making it more obvious!) thanks!

Thank you Leslie for teaching me about that existing function. I tried doing the search, but still no luck. I have added the word "pregnancy" to the narrative section of my profile so that we can see if I appear in the full text search results (this word has nothing to do with my professional identity!). No luck. I also have the same word appearing in my "Faculty Mentoring" section as well. I think this search function isn't very intuitive or easy to spot. And indeed I do think adding the faculty mentoring narrative field for deep text searching would be good. If I'm dreaming big, perhaps having a more sophisticated "advanced search" function that would allow complex queries across multiple fields or domains. Say for example, if a junior faculty wanted to find a prospective career mentor in a specific school (e.g. SOM), with a specific career path that included funding from a particular funding agency (e.g. RWJ), and someone who is balancing family obligations at home. This kind of querying would be most effective when the database is large and one needs to thin down the search results (e.g. when searching PubMed there are lots of mechanisms to limit the results). Thanks!

Hi - thanks for trying the search... indeed, it is not easy to find. Because we are testing this out, we limited the "full text" search results to the first 20 records found. I suppose "pregnancy" is fairly common. I did a search for "Rochester" (maybe someone is looking for alumni :) -- and if you do this, at least you can see the functionality and you will be in the search results! Thanks for the idea and the feedback! It's invaluable for us.

As a result of technical feasibility review conducted by the Virtual Home team at CTSI, this proposal was considered beyond the technical scope for implementing it as an OpenSocial gadget. On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Twitter Widget

Proposal Status: 

I think it'd be useful to add a widget for faculty and staff to share Twitter feeds through Profiles, especially folks who maintain and manage social media across the University.  Perhaps even a way to pull the information at UCSF centrally, a sort of Twitter hub using Profiles as the primary source of data.  

 

I know this functionality already exists to some extent on various sites, but similar to the "co-authors" and "same department" tabs that come up on the right, this could be a way of further integrating concordant research and researchers, perhaps even agregating similar research topics.  For instance, if a researcher chose a certain amount of hashtags to follow, this widget could pull together similar voices at UCSF who are working on the same topic.  

 

It has a clear external value for those trying to navigate the (daunting) institution from Profiles, but also of value to the UCSF faculty or staff as a method of networking.

Comments

This sounds like a great idea, and like it should be discussed with Anirvan and Katja, who won a recent IT Open Proposal with something similar: http://open-proposals.ucsf.edu/itcontest/proposal/139

Thanks, Brian! Yes, especially if there's already some traction around a similar initiative. I think it'd be really useful to use Profiles as a sort of zeitgeist mechanism for social media as a research tool.

We're on the same page! UCSF just developed a simple Twitter gadget — here's an example of it in use. We're in the process of open sourcing the gadget so other universities can use it.

Additionally, CTSI at UCSF has been tracking lists of Twitter accounts of people at UCSF and groups at UCSF. These Twitter lists (perhaps repackaged/filtered) are sort of like what you're asking for, right? How would you want these to be repackaged, so they're most useful for you?

I think something more streamlined of the visible tweets page would be pretty interesting.  Great to hear we're on the same page!  I think the additions I'm suggestion would be fairly simple and straight forward, and a great way to push forward an already exisiting initiative!  

Selected comments from Reviewer(s):

Reviewer 1: "Don't people who share the same hashtag eventually know each other through the twitter streams?  More importantly, I doubt that more than 20% of people in profiles are active twitter users; I could see how this could be valuable in a setting with higher twitter penetration."
Reviewer 2: "depends on how cleanly it's done.  But if it's done already, may not be relevant"

 

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Highlight Selected Publications

Proposal Status: 

Function:  to be able to flag/select top 5-10 publications to showcase at the top, with the complete list included.

Example of what we did at the Dept of Radiology with Grants:

http://www.radiology.ucsf.edu/research/centers/translation-study-osteoar...

Highlighting selected chosen publications, while still maintaining complete list.

 

Impact:  Give researchers ability to show case their 5 top most relevant or most pertinent publications first.

 

Comments

I like this idea. A "Featured Publications" gadget would be simple to build, and help researchers promote their most interesting or relevant work.

Selected comments from Reviewer(s):

Reviewer 1: "Really important innovation that will reinforce the value of the tool in all areas."
Reviewer 2: "...it should just be done. Period."

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Education and Training Field

Proposal Status: 

Feature: Add fields similar to Honors and Awards, but for Education and Training.  Listing the following by dates:

Education and Training
Medical School
Internship
Residency
Medical License
Certificate

 

Impact:  Shares education, training, degrees, experience with similar background.

Comments

I really like the idea, though there's the question of whether it's better implemented as a core product feature, rather than an add-on gadget.

Selected comments from Reviewer(s):

Reviewer 1: "We've wanted this feature ever since we started using Profiles, but the impact is limited.  The benefits are more on the administrative side."
Reviewer 2: "should absolutely be addressed and is important but probably not relevant to this contest."

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Publication Formatting

Proposal Status: 

Formatting proposal:  To format publication entries to resemble similar format as PubMed's more closely, clearly delineating publication title, author, etc. with bold, and underlines.

 

Impact:  Allows easier gleaning of information (title and author) for consistent viewing from profiles to PubMed.

Comments

Did you envision (a) adding PubMed-style formatting to the existing list of publications, or (b) offering a way to export a user's PubMed-formatted publications?

Selected comments from Reviewer(s): "Which format would you select, or would you allow each profilee to select for each publication?  Once in the hands of the profilees, does selecting a format or not mean anything other than the profilee did or didn't have the time to do it? How does selecting a format enhance the research?"

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Awards and Honors Formatting

Proposal Status: 

Formatting issues:  Currently the entry fields for Honors and Awards do not mirror the profile published view.  For example, the entry field for Name comes before Institution, but on the published profile, the Institution is listed first.  I also have to add a hyphen (-) to the beginning of the Name entry, otherwise there is no delineation from Institution and Name of award.  Another formatting inconsistency are the awards listed from past to lasted in edit mode, but reversed when viewed published.

Example: http://profiles.ucsf.edu/ProfileDetails.aspx?Person=4774117

Impact:  These are aesthetics issues, and may make the editing experience a little less jarring.

Comments

This is a great idea. We're in the process of upgrading the Profiles software, and the new system includes an updated edit screen. We'll check to see if the new version of the software incorporates your suggestion.

As a result of technical feasibility review conducted by the Virtual Home team at CTSI, this proposal was considered beyond the technical scope for implementing it as an OpenSocial gadget. On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Allowing Faculty Member to Add Link(s) to Moodle, TICR, or Other Online Forums or Resources

Proposal Status: 

A growing number of faculty members are teaching online. Might it be helpful to allow them to add, to their profiles, a link to their forums, classrooms, blogs or other resources, as hosted in the Moodle/Collaborative Learning Environment, Training in Clinical Research (TICR) or other sites? In some cases (e.g., some Moodle classrooms), the sites will stay open indefinitely to allow the faculty members to use them as "permanent" repositories for videotaped lectures, articles, texts, PowerPoint presentations, and other course material.

I understand that a profilee can add up to five websites to his/her profile. I wonder if adding a section just for online education links would encourage more faculty members to add those specific links.

We would create an algorithm that asks if the researcher has any online learning resources. If so, a second field would appear in which the profile could input his/her own URL(s). The help text (and the “How Profiles Works” page) would offer suggested links, including those I’ve typed above, among others. The default visibility setting for this field would be “hidden” so as not to discriminate against those faculty members who are non-adopters of online education.

This would be inexpensive (or no-cost) and user-directed. It would allow profilees to share their online learning resources more easily with their students, mentees and alums. More importantly, it would also advertise the expanding role of the University on the leading edge of online education.

Comments

Hi - UCSF Profiles already has two open-source gadgets that can serve this purpose. The first, as you note, is the Websites gadget, wherein a profile owner can include any five links they want to present. Profiles also has the Presentations gadget, which uses Slideshare to allow a profile owner to put presentations (or a sample lecture) in their profile.

 

I believe the need is to educate profile owners/online professors to add links and presentations to their profiles more than to build another gadget. What do you think should trigger the algorithm you suggest?

 

Hi,

Thanks for asking. I agree that educating profile owners/online professors would be a crucial step in getting these resources on Profiles. In the interim, might I suggest this method? Please excuse my ignorance of Profiles algorithms and how they work.

Could Profiles integrate with Moodle/the Collaborative Learning Environment (CLE) to find profilees who have active “instructor” status in Moodle? Then an algorithm would auto-populate the Profile for any faculty member with that status, with a link(s) to any of his/her Moodle classroom(s). The Moodle classrooms are password secured and already include PPT’s, videotaped lectures, texts, etc. Ryan Brazell, Brian Warling or the other helpful staff at CLE might be good resources.

I propose this method instead of an algorithm that searches by a keyword search using any of these “exact phrases”: online education, online learning, digital learning, e-learning, flipped, MOOC/massively open online course. This is because I could not find online instructors like Bernie Lo, when I searched by those keywords (“No keyword matches found.”). Using some of those exact phrases, I did find via “full text search”, Bernie Lo. But I also found peripherally relevant profilees like Angel Chen and Leslie Floren, who research or work in online education in general.  Integrating with Moodle might be a more direct and precise way to identify online instructors and provide links to the relevant resources.

Great idea Aria, however, I do agree with Brian that it may come down to more comprehensive education on the part of profile owner/acamdemic. It seems like a more succint and targeted location for this information may reside in owners updating their websites and narratives. How can we better educate the profile owner showing the power of the tool?

Thanks for asking! If we haven't already, might we ask UCSF IT (Elazar Harel, Andrew Riley, etc.) to broadcast the benefits of a robust profile, to a listserv targeting departmental IT administrators, for further dissemination? Such a “top-down” message could educate the profilees about the benefits of updates, and traffic (by students, colleagues, collaborators and other visitors), to their sites. This type of message could also explain how a robust profile facilitates identification of, and outreach to, potential collaborators. For example, UCSF has researchers in dynamic fields like ImplementationScience. Researchers working in this field often don't know others who, or even that they themselves, do relevant work. A “top-down” message with this info might help get the word out. Otherwise, a “bottom-up” message from program coordinators like me  to program faculty, might help.

Selected comments from Reviewer(s):

Reviewer 1: "Not clear that this is even needed, since some of the functionality already exists. Need to plug profiles into more departmental websites to push the need to build and maintain a complete."
Reviewer 2: "education campaign needed… to drive academics to update and use the existing resources in Profiles for education and online learning."

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Sharing and Collaborative highlighting, commenting, annotating of manuscripts and internet full-text articles

Proposal Status: 

Problem. Researchers spend a lot of time reading articles, but it's increasingly difficult to keep up with the explosion of information (volume, velocity, variety). Pubmed and other tools exist for culling articles relevant to our interests, but might generate more results than one has time to read. Often, while reading an article, one might wish to discuss part of it with a colleague or forward it for some interesting detail. However, an emailed pdf or link to an article leaves a recipient with a large amount of matter to read and to further cross-reference with the sender's comments.

 

Suggested Feature, Utility. If highlighting, comments and annotations can be overlaid on, or linked to specific parts of, an internet full-text article- and viewed openly (or by a chosen audience)- that will make it easier and more time-efficient for scientists to share. discover and discuss new information. It will also enable transparent critique of information. Possibly, this system could be used to save the article as a complex of "original article"+Extras/Comments, or print to a pdf file or appropriate digital format. Lastly, scientists will have the option of sharing articles of mutual interest, complete with Comments to their social networks. Or, even solicit comments by linking to the articles from their Profile pages. "Social commenting" can also include "Like" or "Not Like" etc buttons to rate the +Extras/Comments appended to an article.

 

Potential issues. Institutional library subscriptions often dictate whether scientists at any given institution may or may not access the full-text of an article. This idea may initially be implemented on open-access articles, eg. freely accessible via Pubmed.

Abstracts are often always accessible to all, and many institutions share materials via interlibrary loan. So, possibly, the +Extras/Comments electronic versions of articles, or printable files, can still be shared.

Comments

This is an important idea, but doing it well would take a significant amount of work, and might be out of scope for a standalone Profiles/VIVO plugin. Does anyone know of preexisting systems that do this kind of thing, that one could integrate with? Could platforms like Mendeley or Qiqqa meet your needs?

Selected comments from Reviewer(s): "I don't see this as a very useful tool for Profiles. But this technology would be TERRIFIC for all the journal clubs that go on out there with expert opinions/analyses of journal articles that just evaporate into the atmosphere with the JC is done.  If the expert/presenter would use the software to capture their comments and opinions about the article, overlaid on the actual article, then this could be a terrific archived document for sharing and learning."

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

External Collaborator tool.

Proposal Status: 

I would like to propose a tool for representing external collaboration among researchers.  An academic may collaborate with researchers at universities other than their own, or with researchers within industry. Currently these interactions are not represented within a single institution's Vivo or Profiles RNS instance. For example, Vivo and Profiles RNS typically display very few collaborations for a researcher new to an institution, even when the researcher is well-connected.

I propose a gadget that allows a researcher to enter the Vivo RDF URI and collaboration type (e.g., co-author, advisor, or advisee) of an external collaborator.  It would then automatically populate an "External Collaborators" section on the researcher's profile using the collaborator's RDF. This would have the advantage of enabling new researchers to expand their profiles, and it would allow all researchers to provide a more accurate representation of their work.

Additionally, it may be possible to further extend this gadget by adding external collaborations to the triple tables within the research networking tool.  This would serve to add the collaborations to a researcher's RDF, which in turn would be a step towards creating a global distributed VIVO network.

Comments

Hi Nick -- we should discuss this idea.  KNODE essentially has the cross-institution data to make this work.  It would be great to understand your use case and brainstorm about what the gadget might look like.

Steve

I remember talking to you at the VIVO conference. KNODE sounds like a very interesting product, and something that could be very useful to our customers, I'm going to sign up for a Beta Invite on your site, take a look around, and get back to you as i have further ideas / questions about how we could use KNODE, and how it could be used with the external collaboration use case described above.

Nick

Assuming reasonable disambiguation, I think Knode-based gadgets could be an excellent way to provide access to external collaborator data to Profiles or VIVO.

This is a valuable idea and rather imperative as academic research is collaborating more frequently with industry/external institutions. Would these only be academic collaborators? I can forsee an issue if we move into the realm of industry profile integration, but see value in identifying collaborations that are more than academic. Interested to see this develop.

The KNODE system identifies both academic and industry people.  Courtney's point on frequent collaborations across institutions is indeed one of the key motivations behind KNODE.

I would definately like to see a tool for both industry and academic collaborations. The tool I am proposing would work with both, and i believe that the VIVO ontology is descriptive enough to handle both sorts of collaborations, although in reality, few companies have VIVO RDF compatible sites, so linking to collaborators in industry would currently be difficult.

In order to maximize the interoperatability of this tool, it should be built to handle FOAF RDF when VIVO RDF is not available. As the FOAF ontology is a subset of the VIVO ontology, it should be fairly simple to handle FOAF RDF, by displaying just that subset of the information, and given that FOAF is a more widely used ontology, this would enable interoperatability with a much larger number of networking products.

Indeed, this gadget, as well as the Harmonize Profiles and ORCID gadget proposals, would need to "speak" bi-directional RDF to reasonably exchange data between different systems.

Nick - agree, the 'External Collaborators' section could be a nice way for researchers to expand their profile, especially for those new to an institution in which they have co-authored few publications or grants with the researchers there. This could also be valuable for potential collaborators viewing the profile and for mentees looking for someone with external research networks, either academic or industry.

 

As with all most of these gadgets, it seems to me that simplicity for the user is key for adoption. Is the intent that the researcher would "select" the RDF URI from a list as well as the collaboration type?

Rachael,

Thanks for the suggestion. I agree that making the tool easy to use is critical to adoption. I think the best approach to this would be adding an "External Institution Search" to the tool. If the tool was pre-configured with the SPARQL endpoint of other institutions running VIVO or Profiles, it could contain a search capability where you could pick the institution from a drop-down list, then enter the name of your collaborator. The tool could then run a SPARQL query against the external institution, and return the URI of your collaborator.

Selected comments from Reviewer(s):

Reviewer 1: "This could be a nice addition/tool. Would be helpful if the strength/depth of the collaborations/relationships could be represented also… perhaps as simple as weighting according to number of publications together, or grants together…"
Reviewer 2: "v. good idea, doable, very applicable, potential partners - though not convinced that a pure knode plug in is the answer, I think this will have to be developed with room for customization/edits by investigators."

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Suggestions for improving Profiles

Proposal Status: 

1. Extra Titles
Proposal: Added Extra Titles field may be useful for researcher who have multiple titles, and/or do research in multi-displinary fields. Example: Extra titles may include Vice-chair of a committee, PI of labs, Faculty of multiple department.

 

2. Research Interests
Proposal: Add a field for Research Interests. This will help in stating clearly the area of research expertise.

 

3. Expertise/Specialty
Proposal: Add Expertise/Specialty field

 

4. Narrative links capabilities
Proposal: Allow links in the body of Narrative

Comments

These are good suggestions for improvements, however they are related more to the core code and not a gadget or gadgets. (Granted, it's not always clear what pieces are populated by what.)  Also, some of these items are already available and some are coming in the soon-to-be-released upgrade to UCSF Profiles. Following your numbering:

1. Extra titles, where available from the HR data stream that feeds the directory-type data at the top of the profile, are shown. See Dr. Jeff Bluestone or Dr. Deborah Grady.

2. The forthcoming version of Profiles has an "everything" search. So anything entered in a narrative, as Research Interests, or whatever, will be searchable. If we go the route of adding dedicated fields for REsearch Interests, it could be done in the new product' s RDF data structure.

3. Same as #2

4. There is a gadget for links already, we called it "Websites". Granted, it's not in the narrative, which could give some context. But again, this would be a change to the base code - the interface to support links in profile editing and the database. Unfortunately this is not in the forthcoming product, at least not the pre-release versions we've been testing.

Selected comments from Reviewer(s):  "Good ideas.  these are all feature wishes to the core Profiles team already I think. "

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

ORCID gadget

Proposal Status: 

With ORCID launching in October, there is motivation for both VIVO and Profiles to actively support what is hoped to be major community shift to use of unique author IDs.  This gadget could be added to user profile pages to:

 

(a) encourage and facilitate ORCID registration,

 

(b) display that someone has registered an ORCID as well as link to it, and

 

(c) help to keep local and ORCID profiles in sync.

 

There are probably other benefits to attaching an ORCID to a VIVO or Profiles user page in terms of things like connected SEO, web indexing, and facilitating linking of open data in general.

 

See also contest proposal for "Harmonize Profiles Gadget".

Comments

I love the concept behind ORCID and hope that it is wildly successful. Having unique IDs for researchers is way overdue in my opinion. It would solve so many data issues.

I imagine the RDF in Profiles 1.0 could easily support the ID and the gadget would focus more on items (b) and (c).  Please say more about (c). What do you envision the sequence of events to be for synchronization?

ORCID is just getting going, so is not quite yet mature.  It is also not yet widely used.  So in the beginning, what I'd think to have for (c) is the ability to upload publication lists, etc. from Profiles or VIVO up into ORCID to populate that system.  Initially I'd think this would be manually initiated by the user.

 

As this all matures, you can imagine a more automated process -- perhaps email notifications of profile changes that need to be sync'd.

 

One day when ORCID is wildly successful, there may need to be a way to pull publications lists down from that "system of record" into Profiles or VIVO.

The very act of listing an ORCID for every supported researcher could be a useful mechanism to remind researchers that ORCIDs exist. This would work better, of course, if UCSF were to handle the issuance of ORCIDs for all their researchers and having them listed in a central database, rather than making individuals register for them one at a time.

I agree that this idea could be very relevant to the long-term success of research networking systems. 

Another reason why this is interesting is that it could help moving research networking systems into the realm of usage-based/impact-based publication metrics. ImpactStory, one of the leading intitiatives in the field, was one of the ORCID launch partners. In additon, ScienceCard, another tool that helps explore how research impact can be collected and displayed, allows researchers to list all their publications, and the metrics associated with them. ORCID integrates with ScienceCard (free public ORCID API).

This gadget would open up opportunities...

 

Selected comments from Reviewer(s):

Reviewer 1: "good idea, but not sure it will be perceived as a win by our audience, and will clutter (with no obvious value to the end user) an already busy interface"

Reviewer 2: "Awesomely useful, but lower impact than others."

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Citation Integration Widget

Proposal Status: 

The main idea behind the Citation Integration Widget would be to integrate and organize access to citations from publications of VIVO users. The word widget is a bit of an understatement in the amount of work that would be required to program such a large scale platform, but it would consist of the following:

 

A citation template would be used, which fills out the information of a citation form a VIVO user, and leaves the rest to the widget (current date, date accessed, page # and sentence) 

VIVO users who have linked publications would set up a quick citation template.

VIVO users who link or upload future publications would have to do a quick citation template as a requirement.

A VIVO user who wants to use the intelligent or literal content of the publication can do so quickly and seamlessly by the use of hashtags (#) or the use of another integrated function. The system would pull the template automatically and perform the in text citation while also filling in the "resources" section. 

The in text citations, as well as the resources, would be live links to the VIVO user's publication - to the exact line even.

 

The long term objective would be to effectively organize citation information and seek to improve the tedious, traditional, manual method of citation today. This could be spread across many platforms, and its impact would be vast since it is another massive improvement in the quality of how scientific knowledge is shared and used.

Comments

I'm having trouble fully understanding the workflow. Could you share some mock/low-fi screenshots of what this could look like — even a photo of a whiteboard mockup would be helpful. Thanks.

(If you have a picture you want to point people to, you can upload it to a site like imgur.com, and include a link in your proposal.)

I think this is the best way to describe it from the usage point of view:

 

Imagine that my first post, the one you commented on, is a publication.

You are viewing it on my VIVO profile, and the keyword of my publication is "Jack1". In your intext citation, you would write #Jack1-pg32-33 (just as an example) and the intext citation would automatically link to my publication. This would create an in text citation (for example: [12], if it's your 12th source) but would also be an active link to my publication, pages 32 and 33. 

 

The correct citation would be automatically added to your "references" section as well. 

 

I don't think a mockup could be made given that this is just a general concept, but the idea is to be able to quickly source citations and organize them through live links. 

As a result of technical feasibility review conducted by the Virtual Home team at CTSI, this proposal was considered beyond the technical scope for implementing it as an OpenSocial gadget. On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.

Using Twitter data in combination with search results

Proposal Status: 

This gadget is designed to use the Twitter API and RDF data from search results conducted on the UCSF Profiles site to return relevant data about search. Using the query terms, the gadget will pull relevant data from currently trending tweets and display these results along with the search results. We would like to be able to highlight the differences in tweet content so that tweets featuring links to articles, resources, and conferences are given priority. 

Comments

This wouldn't be difficult to build. For me, the question is how to filter tweets so they're the most useful.

For example, I searched for "lung cancer" on Twitter, and here's the top 5 results:

Are these the kind of results you were thinking of?

I think it would be interesting to 'highlight' tweets that are relevent to the logged in user.  This could be done by creating a relvancy score between any external content (a tweet) and a person (the logged in user) based on the persons RDF.  

For example, if a tweet comes by that mentions 'stroke' and the logged in user has 'stroke' listed as a vivo:freetextKeyword, or in their vivo:overview, then that tweet would be highlighted.

Ideally we would create a general purpose relevance score that would show how relevant some arbitrary content is to a person.  I can see how this could be done via some search engine components and a weighting being applied to properties of the vivo ontology.

I think that this is an interesting idea using Twitter as a resource for research-related content (e.g., articles, resources, and conferences).

I agree with Eric and Anirvan that it would be useful to filter the tweets in a way that matches the content with user interest such as research topics (UCSF Profiles keywords) or specific disease areas. 

I'd like to learn a bit more about how trending topics would be used here. It is often hard to understand what they are about. However, a high percentage of trending topics are hashtags, so it might be helpful to leverage healthcare and disease hashtags for this project idea (http://www.symplur.com/healthcare-hashtags/regular/; http://www.symplur...).

The following research paper might also be interesting: "Twitter Trending Topic Classification" (Link http://cucis.ece.northwestern.edu/publications/pdf/LeePal11.pdf) The authors looked at Twitter trending topics and categorized them into general categories (e.g., sports, politics, technology). They conclude that a network-based classifier performed significantly better than text-based classifier on their dataset. 

 

 

 

 

 

 

Selected comments from Reviewer(s):

Reviewer 1: "More relevant to desk researchers, I suspect. Unmoderated data means the relevance of the tweets would be up to the reader to determine."
Reviewer 2: "big effort to do it well I think.  Also, one thing to think about is where do inserts/gadgets make sense from a single investigator viewpoint, or from a thematic viewpoint.  This may be more relevant for those exploring the concept search in profiles 1.0 where the twitter content maps onto the stream of recent pubs on a topic from UCSF..."

On behalf of Clinical and Translational Science Institute at UCSF, thank you for participating in this contest.

Commenting is closed.