/
2017 User Feedback and GOBii Responses to Comments

2017 User Feedback and GOBii Responses to Comments

Created by Star Yanxin Gao, last modified on Sep 24, 2018

In July 2017 a survey was sent to CG users to gauge user acceptance and satisfaction feedback to the GOBII system and project management. The survey included eight categories, averaging four questions in each category, and were circulated within GOBII teams at CIMMYT, ICRISAT, and IRRI.  The categories were;

  1. Communications and engagement

  2. GOBII Data Extractor User Interface

  3. GOBII Data Loader User Interface

  4. GOBII Application Tools

  5. Requirements gathering process

  6. Summary for GOBII Ithaca team

  7. System administration

  8. Trainings and workshops

Ratings were from 1 to 5 where 1=lowest and 5=highest satisfaction score.

 

Summary

There was a high overall satisfaction with GOBII with an average score of 4.03. The highest satisfaction rating was for GOBII Trainings and Workshops. Most room for improvement stemmed from the GOBII Data Extract User Interface, which users said could be more intuitive and less restrictive.  In response, we are redesigning the extractor to be easier to use based on specific feedback collected at the GOBII Annual Meeting in August 2017. Some positive comments collected during the survey were;

  •  “The Cornell team promptly answers every email and request”

  • “GOBII developers’ technical competence and their friendly, cheerful, energetic support is greatly appreciated”

  • “JIRA is efficient for issue-tracking and communication”

  • “GOBII Overall is doing exceptionally well, being successful on all fronts! Congratulations!”

 

Figure 1. shows the specific questions asked during the survey within each category, with the solid yellow line showing average rating responses, and the dotted yellow lines showing the standard deviation of the responses. Responses are split according to institute where the survey was carried out. No bar indicates that the question was not answered at that institute.



Table 1. Comments provided by users for each category of questions. Comments in green are those that users considered to be improvements needed, and that have or will be adopted by the GOBII team. The remaining comments are under consideration and implementation will be dependent on the availability of people (time) and financial resources.

Section

Achievements

Improvement Needed

GOBII response to ‘Improvements Needed’ comments in Jan 2018

Communication and Engagement: Meetings and Communication Tools

  • Meetings provide a good forum to understand the current status and upcoming advancements within and across institutions

  • The Cornell GOBII team takes very good notes in online meetings and makes them available on Confluence

  • JIRA is efficient for issue-tracking and communication (though not everyone has access to it via the Cornell VPN)

  • HipChat has enabled some quick feedback from Cornell team members

  • Zoom seems to be working as well as WebEx

  • In-person meetings are expensive but very productive

    • Set up a mechanism to send reminders of action item assignments a week or two after each meeting

    • Develop a strategy to promote inter-Center awareness of bug reports, bug fixes, and feature requests in JIRA without flooding in-boxes with email notifications

      • Regular review at App team meetings

      • Weekly email digest

      • Dedicated space in JIRA (managed by new QA/QC person)

    • Review Confluence “Spaces” and describe changes in organizational structure with Teams

 

  • Action items are now sent out as reminders ahead of meetings

  • We will set up a weekly or biweekly report of bugs reported with links to tickets being processed in JIRA

  • We will add bug reviews to Application and Dev Team meetings

  • We will follow up with Applications team on how best to re-organize Confluence – we agree it needs to be cleaned up

Release timelines, onsite visits, hackathons, workshops, meetings and trainings

  • The Cornell GOBII team has been very good about trying to estimate realistic release timelines and about communicating changes to development and release timelines to all Centers

  • Hackathons and workshops are very useful

  • A lot of care is taken that workshops and other meetings fit well for everyone and are well-planned

 

  • Include systems administrators in hackathons to align system deployment with development teams.

  • Consider modifying the structure of curator trainings to allow more time for coordinated testing of new versions of the system – perhaps by organizing “pre-meetings” and/or “post-meetings” at all Centers

 

  • We will start to include sys-admins more in communications and trainings. We will re-initiate GOBII sys-admin team meetings

  • We will initiate monthly curator trainings to test new functionality and test loading with each CG center

Requirement Process

  • The Dev team does a good job of providing time estimates on the Requirements so that Centers can factor that into their prioritization for each feature

  • The Cornell team seems to do the bulk of the work on getting the requirements documented. That is a lot of work!

  • The Cornell team is always very open to our comments and suggestions and asks for clarifications when necessary

 

  • Slightly longer descriptions of some requirements on the Excel survey could help diverse contributors participate in the prioritization

  • It might be helpful to group together requirements that have dependencies, e.g. a haplotype loader is required to enable haplotype extraction

  • Sometimes it is hard to gather the next round of requirements, especially from "users," when they have not seen the current system and do not know what it can do and what it cannot do

 

  • We will try and be more clear and provide better descriptions for new requirements

  • We will group together requirements with dependencies and be more clear what the dependencies are

  • We have moved to shorter release cycles of 3-4 months so users should be more familiar with the latest features and be better able to prioritize the next round of requirements

Training and Workshops

  • On-site visits are good and are very needed

  • We are very grateful for the help of the Cornell GOBII team in setting the agendas, running the meetings, and taking great notes during on-site visits and trainings

  • The GOBII team is wonderful about sharing their presentations for future use

  • We appreciate the on-line availability of the Cornell team to help when we organize internal test sessions and trainings

 

  • More frequent on-site visits could be beneficial

  • Targeted presentations to different user groups could increase engagement

  • Staggering the arrival of Dev and App team members during deployments could provide a better chance for the App team to show a functional database to users

 

  • We are increasing our training beyond the GOBII team – reaching out to other CG centers and NARS. We are adding GOBII trainings on adjacent days. We will increase our developer on-site visits

  • We will have focused curator meetings. We are developing focused training modules for outward facing training

  •  We are now deploying new versions of GOBII remotely and so the development and applications team can engage at staggered times

 

System Administration

  • The v.1 deployment process was faster than previous deployments

  • The Dev team came to visit each Center during the v.1 deployment to provide on-the-ground support

  • The system can now work with multiple authentication methods required by the different Centers

 

  • More documentation for Sys Admins, including information on how to make changes in the container persistent would be helpful

  • Sys Admins would benefit from start, restart, stop scripts for the system and scheduler support for the analytical components (nodes)

  • Some additional planning is required to have a robust pipeline for data loading and updating on “test” and “production” systems for the short term (no delete or update functions) and for the long term

  • Improved data “security” options could be introduced to (1) Warn users about sensitive data usage (2) Implement user groups to limit access to files so that everyone cannot see everything

  • Functionally differentiate roles like PI and data curator

 

  • Sys admin documentation has been greatly improved and will be made public to improve access. Sys-admin teams will be better engaged in general in the future.

  • Sys-admins meetings will be used to address the needs of sys-admin and improve communication.

  • Moving data from test to production can be addressed in sys-admin meetings

  • A post-configuration tool is being developed that will allow sys-admins to be able to start and stop the system

  • We now have a warning on the UI and in emails that data is sensitive

  • We will discuss how to limit files to particular users but this will be a major enhancement

  • Functional roles is on our list of enhancements needed but has not been prioritized highly by the team

Data loading

  • Straightforward process

  • Nice ability to flow from Project down to Dataset in the loader

  • The email messages improved greatly from the first version of the loader

  • The saved templates for loading data are useful

 

  • Add delete (and eventually update) functionalities to enable data loading to a production database

  • Fix error message bugs, e.g. report dataset import failure even when it succeeded (v.1.1 J)

  • Remove need to login each time a dataset, marker file, or sample file is previewed before loading

  • Enable some form of review of template mappings – in general or prior to loading

 

  • We are working on ‘soft delete‘ loader UI functionality and a sys-admin ‘hard delete’ tool

  • Error messages have been greatly improved for loading and extract

  • Extra login is being removed and will be released soon

  • Updates to the data loader will enable review of template mappings

  • We will generate standard data templates for high volume vendors in collaboration with the HTPG project

 

Data extractor

  • The extraction works very quickly

  • Emails contain the exact path to the download folder

  • Users can now subset by markers or by samples/germplasm

 

 

 

GOBII Core Features

  • GOBII so far did a very good job in linking in with other projects, initiatives, such as the HTPG and EIB. This is very much appreciated as integration of GOBII is going to be really important.

 

 

 

Connected Tools and Pipelines

  • Flapjack MABC:

  • Quick development and continual enhancement of a useful practical tool

  • Can potentially save breeders significant amounts of time.

  • Flapjack F1 QC: In conjunction with the HTPG, can facilitate the routine incorporation of QC into breeding pipelines

  • GS pipeline

  • Galaxy wrapper works

  • Significant progress was made quickly at the hackathon

 

  • Flapjack MABC:

    • Enable the generation of the QTL files directly from GOBII in the future

  • GS pipeline:

    • Create a Galaxy wrapper for the GOBII extractor

    • Work to connect to other systems to the processing of the genotypic and phenotypic data

 

  • We are working on requirements to allow for new QTL formats in Flapjack that can be supported by GOBII

  • We have developed a Galaxy wrapper to allow for GOBII data extraction, but to implement this will require prioritization and in the meantime we will likely use a shared data folder system to integrate GOBII and Galaxy

  • We are working with the new BrAPI coordinator and Excellence in Breeding module 5 to connect systems via APIs

LikeBe the first to like this

Write a comment…