I. Expertise Locator [Federal Knowledge Management Working Group (KMWG)]


I. Expertise Locator

Overview

Let's use this space to share your experiences on developing an expertise locator. Tell us your story. All stories and ideas are welcome -- success, failures, and more such as what are the key requirements an experts system must include.

Click the Discussion Tab above to contribute to the conversation.

Case Studies

NASA has created an expertise locator that does not require people to maintain profiles. The work of Andy Schain, Kendall Clark, and Mike Grove, POPS (people, organizations, projects, and skills) allows NASA employees to find experts based on those things that define expertise, rather than only a self-asserted competency. It builds on a technological bed of semantic technologies (including Jspace, RDF, and OWL) and algorithms for weighting the importance of different aspects of a person's history. It reaches to NASA's publication database, workforce system, competency management, and organizational information to infer expertise from the work that is occuring and the outputs of the work. It is sponsored by the NASA Office of the Chief Engineer. You can follow up with Jeanne Holm or see more on POPS at:

http://www.w3.org/2001/sw/sweo/public/UseCases/Nasa/
(If you have difficulty accessing the above link, click the Discussion tab for this Wiki page and you'll find a link to a PDF of the Case Study.)

Demos of the NASA and DATA.GOV use cases can also be found at http://clarkparsia.com/pelorus

Tags : Experts, Storytelling, Brainstorming Requirements

Short Link: http://wiki.nasa.gov/cm/wiki/?id=7256

Posted by Brand Niemann on Oct 28, 2010 1:51 AM | Reply To This | Permalink

Please see my pilot demo for Expertise Location at http://federaldata.wik.is/Expertise_Location

Posted by Ellie Trevarthen on Oct 20, 2010 2:55 PM | Reply To This | Permalink

Subject Matter Expertise Identification System

Posted from the KMGOV listerv

Comment by Paul Villano:

I'm still trying to establish a Subject Matter Expertise identification system. (I've raised this before but it was lost in the miasma somewhere) Processes, procedures, techniques, tools, best practices? Anybody?

Also we are re-emphasising Moral Leadership Training. Definition: Training on the virtues and values that were formative in the shaping of America and are still present in the contemporary military setting. This training recognizes the inherent dignity of all people, the value of the state, the virtues of leadership, selfless citizenship, and duty. Moral leadership training also examines the religious and spiritual connections associated with ethical decision making, personal values, and personal relationships.

Any resources & ideas for collaboration and information gathering you can provide would be appreciated.

Posted by Ellie Trevarthen on Oct 20, 2010 3:34 PM | Reply To This | Permalink

RE: Subject Matter Expertise Identification System

Posted from the KMGOV listserv

Comment by Peter Yim

For those who are interested, Ontolog ran an exercise and a demo, where more than 8 years' worth of transactions on the [ontolog-forum] mailing list were semantically parsed, making it possible for, among other things, an "interest and expertise" search
and matching. This was done in collaboration with the OASIS ICOM TC (Integrated Collaboration Object Model), Oracle, DERI and others, and presented as part of the session on "OWL 2: Tools & Applications."

Check out the archived session at:
http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2010_08_05

... and, in particular,
http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2010_08_05#nid2FC6

(where details on how that was implemented was described.)

Posted by Ellie Trevarthen on Oct 20, 2010 3:31 PM | Reply To This | Permalink

RE: Subject Matter Expertise Identification System

Posted from the KMGOV listserv

Comment by Joyce M. Hunter

There are a few best practices in the state tax and commercial financial space that offers real-time data matching that has been very successful. Let me know if anyone is interested in the details.

Posted by Ellie Trevarthen on Oct 20, 2010 3:01 PM | Reply To This | Permalink

RE: Subject Matter Expertise Identification System

Posted from the KMGOV listserv

Comment by Jeanne Holm:

We've had a lot of luck with using a semantic technology solution that infers expertise from what we already know about our employees and partners (publications, workforce planning, training records, etc.), rather than having them all create profiles. Profiles tend to get out of date and are often under- or over-inflated. Just at NASA, we save ~ $30M annually from not having to require each person to create and maintain a profile.

The system at NASA is called POPS (People, Organizations, Projects, and Skills) and is part of our KM work. Richard Detsch has brought up a similar capability at the Army Research Labs. I'll be extending this work to the Air Force and Navy this year as well.

There's a good technical description of POPS at the W3C site: http://www.w3.org/2001/sw/sweo/public/UseCases/Nasa/

Posted by Ellie Trevarthen on Oct 20, 2010 3:29 PM | Reply To This | Permalink

RE: Subject Matter Expertise Identification System

Posted from the KMGOV listserv

Comment by Chad Hinsch

I am interested as well. Just yesterday I had a COP leader come ask me for advice on how to set up a system that would allow people to quickly find others with experience relevant to their problems. Our organization has lots of examples of what doesn't work (Like Jeannie said, having people maintain their profiles is not feasible in the long term), but zero knowledge on a system that works.

Posted by Ellie Trevarthen on Oct 20, 2010 3:06 PM | Reply To This | Permalink

RE: RE: Subject Matter Expertise Identification System

Posted from the KMGOV listserv

Comment by Brand Niemann

Is the POPS database available for reuse? I would like to try some new semantic technology on it just like I have recently with data.gov.

Would be happy to talk with you about the SME identification and locator. If there is a broader interest in this topic, perhaps we could begin to build out a body of knowledge and experience around it.

Posted by Ellie Trevarthen on Jul 28, 2009 12:33 PM | Reply To This | Permalink

Chat Stream Q&A -- 7/28/09 POPS Expertise Locator Demo

NOTE: Presentation from this demo is at: http://wiki.nasa.gov/cm/wiki/?id=7256

Someone asked: What is LDAP?

Ellie Trevarthen: Lightweight Directory Access Protocol

Ellie Trevarthen: http://en.wikipedia.org/?wiki/Lightweight_Directo?ry_Access_Protocol

Jeanne Holm: lightweight directory access protocol - LDAP, think of your agency directory

Bill Neufeld: What kind of "personnel information"

Kendall Clark: items such as email address, office, phone number, department number, etc.

Patricia Eng: No PII? IS degree data PII?

Neil Olonoff: Generally speaking, what are the requirements for the source data?

Kendall Clark: in addition to what they're good at, what they've worked on, and so on

Kendall Clark: No, no PII is used in POPS

Jeanne Holm: We exclude PII from the index we build

Jeanne Holm: Requirements for source data include that it is stewarded, authoritative, up to date, and they will sign a service level agreement

Kendall Clark: Neil, the requirements are loose, most semi-structured or strutured source will work, such as excel files or RDBMS systems

Bill Neufeld: how does the what they worked on, what they're good at get into a personnel record? Or is there more than just the officical personnel record?

Patricia Eng: This doesn't include contractors, does it? Or details or rotations.

Kendall Clark: Bill, the data that goes into POPS comes from multiple sources, so what they've worked on comes from one source, and what they're good at comes from another

Sam: How is this different from yahoo et al ca. 1996/97?

David Heiser: Does JSpace have any security issues?

Jeanne Holm: What they worked on is part of our workforce management system (what are you charging time to and been assigned to)

Kendall Clark: Patricia, no, no contractors are included right now

Kendall Clark: Sam, I'm not sure I understand your question

Jeanne Holm: It includes contractor or rotational information in that they publish presentations and papers that are represented in the data

Kendall Clark: David, what kind of security issues do you have in mind?

Patricia Eng: I worked at NASA as a detailee. SO my expertise would be overlooked and perhaps lead to hiring yet another contractor to do the same job.....I was not asked to do formal presentations or papers. Everything was done in house and to teh best of my knowledge not captured other than in the SMC meeting minutes.

Bill Neufeld: The taxonomy is derived from the meta data, address, e-mail, last name, etc.?

David Heiser: ACL issues, FISMA certification, and the like.

Kendall Clark: David, only registered users can get to POPS to see the data, there's not specific ACL within the whole dataset, but we could put ACLs on the parts of the data

Kendall Clark: Bill, taxonomy is derived from the data sources, yes

Anthony Arrington: Will Data Management/Data Change Management be covered in this presentation?

David Heiser: What is SPARQL?

Kendall Clark: Patricia, yes, this is an issue with the sources, if you don't end up in the original sources, we can't find you. I'm a contractor and I face the same issue

Kendall Clark: David, it's a query language similar to SQL

Kendall Clark: Anthony, not really. Data management is handled by the data owners

David Heiser: I don't know if my fed arch will accept SPARQL -- is it part of the federal enterprise standards?

Kendall Clark: David, I'm not sure if it is, But if your architecture supports SQL, we can convert the SPARQL into SQL

sam: DH: Most modern database systems will support SPARQL.

Kendall Clark: There would need to be a small piece of middle ware to do that conversion, but it's a trivial setup

Kendall Clark: Oracle has very nice support for SPARQL, fwiw

sam: Have a look at http://www.w3.org/2006/Talks/0301-melton-query-langs.pdf if you're curious about it.

Addison - GAO: Which ETL tool was used?

Kendall Clark: Clover with some custom plugins to handle RDF

Kendall Clark: I believe its cloveretl.org

David Heiser: How do you determine which data sources to query to determine the expertise?

Patricia Eng: What is ETL?

Kendall Clark: ETL is an acronym for Extract Transform and Load -- converting data from one place to another

Kendall Clark: David, that's usually up to the organization and what they ultimately want to use jSpace to browse

Kendall Clark: So in the case of POPS, since it was for expertise location, NASA chose sources relevant to personnel information

David Heiser: If organizations consider their data sources proprietary, I might need a big hammer to encourage participation, otherwise the locator may be inaccurate.

Kendall Clark: David, another factor in choose the sources are criteria such as authority, currency, access, etc.

Patricia Eng: I think this is for an agency only - not govt wide - right?

Kendall Clark: Yes, POPS is agency only

Patricia Eng: So there is no way to determine whether the person is competent?

Bill Neufeld: In the details window, is that from the personnel work?

Steve Martin: How exactly is the info updated?

Patricia Eng: There are a lot of people that work in organizations, but not all are good or competent. How do you sort out the good ones from the others?

Patricia Eng: But you only capture the NASA employees not the contractors, even if they have been there for years....

Steve Martin: Are there any union or bargaining issues with publishing this information? if yes how were they overcome / delat with?

Scott Petrie: There used to be a way to download and play with jspace, is that still available?

Patricia Eng: Can someone give us the contact info for the presenters?

Patricia Eng: Also, is the licensed for government wide or is this only for NASA?

Addison - GAO : How are POPS and jSpace licensed exactly?

Kendall Clark: Patricia, Yes it's licensed gov't wide, there's a slide about that in the presentation, which we'll provide

David Heiser: How much to develop?

Kendall Clark: Just for reference, you can contact us at kendall@clarkparsia.com

Kendall Clark: You can get a map of our office location on our site, http://clarkparsia.com

David Heiser: Is this technology that NASA can share with other government agencies?

Ray Chance: David - licensing is free to all govt agencies

Ray Chance: Requires service to integrate with your systems

David Heiser: How long did it take you to build the NASA application?

Kendall Clark: It took a few months to build the tool originally

David Heiser: How do you maintain the relevancy of the data?

Kendall Clark: David, the data owners are responsible for maintaining the data

Anthony Arrington: Are you able to ask questions in the system?

Patricia Eng: So the current system is based on existing NASA direct staff data?

Kendall Clark: Yes

Kendall Clark: Anthony, what kind of questions?

David Heiser: How often is a refresh done? Or is it a constant query?

Steve Martin: Have you run into any roadblocks from union or bargaining units with sharing an expertise directory? if yes how overcome?

Addison - GAO : You could add jSpace and POPS to the GOTS page on Wikipedia

Patricia Eng: Which is limited in that it doesn't include contractors or verify the competency levels of staff other than input from the employee?

Kendall Clark: Patricia, yes it's based on direct NASA data

Anthony Arrington: Questions such as, " Who has worked on methane fuel projects for the Southwest region as a program manager in the last 6 months?

Anthony Arrington: Yes, plain text.

Patricia Eng: What about asking about a specific type of expertise -- say heat shield technology?

Kendall Clark: Patricia, if that level and type of expertise is included in the original source, CMS in the case of POPS, then yes

Richard.Detsch: How are research papers handled in terms of entering search criteria?

Kendall Clark: Papers (in POPS) have two facets associated with them, author and topic. So you can find papers by author and/or by topic or subject of the paper

Posted by Ellie Trevarthen on Mar 30, 2009 12:37 PM | Reply To This | Permalink

Expertise (aka SME) Locator

I just tried and was able to access http://www.w3.org/2001/sw/sweo/public/UseCases/Nasa/, but in case others have a similar difficulty, attached is a PDF of the Case Study.

Ellie Trevarthen
Facilitator, KMWG Wiki and Listserv

Attachments

File	Size(Kb)
POPS_Case_Study.pdf	468

Posted by Jack Lapke on Mar 30, 2009 11:17 AM | Reply To This | Permalink

Expertise (aka SME) Locator

I'd like to learn more about this, but I couldn't access: http://www.w3.org/2001/sw/sweo/public/UseCases/Nasa/

A while ago we tried a spreadsheet subject matter expertise (SME) locator and as expected it didn't work because it required members to manually enter and maintain their desired SME areas.

We use SharePoint portal so the search function can be used to maybe find people with desired expertise. We also encourage members to maintain their "My Site" with skills details. Some do it and some don't.

- Jack Lapke

You are not allowed to comment on this Page.