Product & Timeline

Roadmap Canvas:

This is our notepad for where we can record themes, high-level user outcomes, epics and stories.

https://miro.com/app/board/o9J_knTHCtM=/?moveToWidget=3074457349556854726&cot=12

Product Feature Ideas

List of features based on the HMW solutioning exercise

In Scope

Bulk Upload UI improvements

• Apps supported

• Autodetect

  • Naming conventions

• CSV Upload for uploading and importing files only, not for batch analysis

Batch analysis

• Shared parameters for single bulk job

• Apps supported

Original Scope Presentation

Data upload improvements:

• Focused on reads, assemblies and genomes (a.k.a. MAGs)

Target use cases:

• Kelly Wrighton WHONDRS data + analysis

• ENIGMA data + analysis

This project will only allow for *sets

• Homogenous set types

• Support Homogenous groups of objects or *Sets associated with SampleSets can be used as well

List of apps and data types in scope to support:

Affected Apps and Types Scope: https://docs.google.com/spreadsheets/d/1MoVfXBQVzvbUI_LgFuF_-xQ1d0dQUbTY2g4oyohRfIg/edit#gid=0

Reference: KBase's own list of apps overall, with a deprecation tab and current list (might not be complete) https://docs.google.com/spreadsheets/d/1FHCWkqrT5J2FgCj1pF-KVRh366iQGWzyWHGN-s35lAI/edit#gid=1903256230

Out of Scope

• No new analysis tools, pipelines

• No support for JGI MAG data

• No processing of new data types not already handled

• No mixed data type processing within a set

• No technical work by this team on implementing SampleSets (creation/editing, filtering, viewing or GUIs for subselecting for an app running)

• Data palettes

• Import from Web apps

• Running multiple parameters for the same bulk app

• Redesigning the Import UI (vs. improving the UX through an app cell)

• Allowing the user to order specific files to have analysis done first, or applying ordering in the queue

Project Outcomes

• A repeatable model for product management, design and software development best practices customized to the specific needs of KBase

• Artifacts detailing the iterative software development methodologies used on the KBase project to be used as a future model for KBase team’s continued work in iterative software design beyond the scope of this project..

• Received assistance and training in the process of delivering:

  • A prototype to fulfill the highest-priority use case - uploading datasets into the KBase system so the user can easily upload compendia (collection of organized related data of a single type) and perform meaningful analyses to generate high-value data products from that data.

  • Infrastructure changes that enable upload and batch operations which serve Knowledge Engine and Concierge needs.

Success Metrics

Scoping Document for this project that we worked on a bit with them. Success measures are not complete: https://docs.google.com/document/d/1MduFTlesTQSjkDNfWnJ8Wu0tB-o1G2BFuRTsKbhtPkQ/edit?ts=5ec2c33f#heading=h.340xn1cjmpgi

Initial Success Measures: [updated 8/27/20]

1. Improve the bulk upload user experience

   • Measured by:

     • Getting positive feedback from internal stakeholders including KBase staff, JMC, Janaka, Lauren Lui

     • Getting positive feedback from external stakeholders

     • Can get a quantitative assessment, like scale from 1-5--use as a relative measure, comparing iterations (not an absolute measure)

     • Reduce effort user needs to do bulk upload and import [TBD #]

2. Users are able to upload 10s-100s of data objects in a single action using bulk upload

   • Measured by:

     • Lower error rate per active users for the month uploading 10-100s of data objects in a single action than pre-release (may need to wait until there's enough volume on the system post-release to measure)

       • (accounting for fewer users)

     • Lower # of Helpdesk tickets (%TBD)

       • Tie to error rates

     • Error recovery - people find it hard to recover from errors without asking for help

     • (TBD %) increase in data being uploaded overall

     • Getting positive feedback on increase in data being uploaded for some users (revisit user interviewees)

   • Adoption - how to measure usage

3. Analysis: Users can run analysis to match the WHONDRS data analysis narrative, scaled to 10s to 100s of objects

• Measured by:

  • Users able to run analysis on the WHONDRS narrative workflow on multiple objects (10+)

  • Users able to run analysis on the ENIGMA narrative workflow on multiple objects (TBD)

  • Users find that the bulk analysis user experience is improved since pre-release [interviews with stakeholders, template for the future]

  • Users can do bulk actions without having to use code cells (having that tech skill)

Initial hypotheses:

• 1. If we improve the bulk data import user experience and run analysis-oriented KBase apps on a set of commonly used data types--users will upload more data to be shared on the system

  2. If we improve the bulk data import user experience and run analysis-oriented KBase apps on a set of commonly used data types--groups will be more likely to bring high-value data into KBase because they will be more self sufficient

[ENIGMA and WHONDRS data without involving KBase staff to upload]

DRAFT Rollout Plan Target Dates

November/December:

• Alpha: Bulk Upload for internal Users

  • KBase staff

  • Janaka, JMC, ENIGMA group

January/February 2021:

• Alpha: Bulk Analysis for internal users

  • KBase staff

  • Janaka, JMC, ENIGMA group

• Beta: Bulk Upload for testing with partners

March/April 2021:

• Beta: Bulk Upload testing with Kelly Wrighton’s group, JGI

• Beta: Bulk Analysis testing with Enigma and Kelly Wrighton’s group