Child pages
  • 202010- XSEDE Monthly Workshop Longitudinal Study Data Sources
Skip to end of metadata
Go to start of metadata

Decisions:

 Summary

Description

  





Action Items:

Summary
Description
Responsible
Due Date

Provide training data QMA-421 - Getting issue details... STATUS

Tom to provide list of institutions and contact list

Tom Maiden




 

 








 

Notes/ Discussion items:

Study 1—Training: Nearly 6,000 completers of the XSEDE 2.0 HPC Monthly Workshop will be followed in the XUP database to ascertain their demographics, employment status, and engagement with XSEDE before and after workshop participation. They will also be asked to complete an online survey to assess the extent to which the workshop affected content knowledge, skills, job performance, and other variables. We will also follow up with institutional representatives to assess the institutional impact of the course.

Retreat Notes:

  • Eval team lead(s): Lorna

  • Eval team deputy: Tonya

  • Eval team deputy: Claudia

  • March 2020 Quarterly: 

  • Evolution of multi site deployment
  • April 2020: Maytal/Lorna retreat (TBD due to COVID-19)


Proposed Timeline:

March 2020 - Quarterly: Longitudinal studies update (Training, KPIs, ECSS data sources).

April 2020 - Maytal/Lorna meeting re: portal training data

May 2020 - Training data set

June 2020 - Query database for Trainee and Broadening Participation (Individual) data - CEISMC interns - Maytal 

June 2021 - Training presentation at June Quarterly

July 2021 - End HPC Monthly Workshop Training Study


Details:

(Notes from meeting):

Goals:

 

Fill any gaps

 XUP data

-       Demographics

  • May need to use only trainees who are active on the portal
  • Oversample, assume a certain amount of failure
  • XSEDE may need to send a reminder email to update data
  • Send everyone an email, see how much failure we get
  • Same user might have multiple accounts because they move around
  • Not typical XSEDE users, may just sign up for workshop

-       Employment status (probably doesn’t update if it changes, same for email addresses)

-       Training data

-       Engagement with XSEDE before/after workshop

-       Data on multiple attempts at same class, users taking multiple trainings

-       Question users: Did they learn from it, what did that enable them to do (Ind. and Inst.)

-       Campuses that use XSEDE as training options: Delaware, Houston Clear Lake (Tom to provide list of institutions and contact list)

-       Which institutions participate a lot?

-       Get post-training testimonials from Tom

-       What kinds of institutions are using trainings? Value? (National labs, NASA, MSIs, EpSCor, etc.)

 Historical post-workshop surveys

-       Demographics

-       Employment status at workshop

 Online follow up survey

-       Measure content knowledge, skills, job performance

-       Ensure survey collects data missing from XUP and historical post-workshop surveys

-       Usage of resources in/out of XSEDE

-       Publications

-       Use of new skill: immediate or later? Do they continue to use the skill

-       Was the training enough to give them the information they need to do what they are trying to do

-       How critical is this training? Are they just curious about the topic?

-       Would you refer your colleagues to this training? DID you refer your colleagues to this training (some time later)?

 

 Interviews with institutional representatives

-       Campuses using trainings strategically

 

Interview questions

-       Different ways institutions use the trainings (staff training vs future user)

-       Are there topics that sites focus on at the exclusion of the others? Does the training fill a gap? Have the gaps changed over time?

-       Some trainings impact institutions outside of the US

-       EpSCor and MSI institutions – check Carnegie list

-       Is there an opportunity to formally integrate the courses with their institution?


Philosophy of trainings: Building blocks of HPC, not bleeding edge of technology

  

Selection will be from ALL events (multiple use trainees only get one email)

Registered vs attended data

 

Positive outcomes:

How many trainees went on to use allocations/HPC resources with XSEDE? (Tom says “incredibly useful”) Match databases

Any publications?

Educational allocations

Research allocations

National labs – did you go on to use HPC resources anywhere?

Big data/deep learning trainings – students use content on cloud, not HPC (eg, XSEDE platforms, DOE labs)

Students take the course, decide it’s not for them (still useful to know)

Diverse objectives from students, diverse pool of student sources (broadening pool)

  

 

All trainings are open-source and transferable. Trainees are encouraged to take it back to their institutions. Students don’t have to use XSEDE.

Trainees log in during the session, training accounts are available if student forgets password, etc.

New grant for every class, Tom and John are PI

 

Concerns about deployment (if emails are out of date, for example)

 

New workshop (hopeful launch late summer): Software engineering (GitHub skills, etc.)

 

Summer bootcamp: 4-day immersive event

 

 

  • No labels