1 of 21

Operationalizing

Better science in less time:

Lessons from the Ocean Health Index

Julia Stewart Lowndes, PhD and the OHI team

Openscapes Director, Marine Data Scientist, Senior Fellow

National Center for Ecological Analysis and Synthesis (NCEAS)

University of California at Santa Barbara

lowndes@nceas.ucsb.edu; @juliesquid

slides: openscapes.org/media. CC-BY 4.0

Artwork by Allison Horst for Openscapes

2 of 21

  • Environmental assessments
  • Open data science
  • Empowerment, resilience, and inclusion

This is a story about...

3 of 21

  • Need for science- and data-driven methods to measure what people care about
  • Need for standardized but flexible methods to assess different geographies
  • Need to streamline assessments from year-to-year to track change through time

The Ocean Health Index

A scientific assessment method, tool, and community for channeling the best available scientific information into marine policy.

4 of 21

Upgrading was possible because of open data science: the tooling and people enabling reproducible, transparent, and inclusive practices for data analysis

OHI’s entryway to open science was through data science

Because our team struggled to efficiently build from our own work

It’s enabled us to efficiently lead annual assessments globally (2012-2020+) & smaller scales (Northeast, BC), and support 20+ independent groups (OHI+)

5 of 21

Open data science is like the Force

More powerful than you ever imagined, broadens the scope of challenges you can tackle

6 of 21

Open science requires teamwork

Inclusive teams with diverse backgrounds can powerfully leverage open data science

7 of 21

Tooling

tools & practices

People

teams & community

Open data science

8 of 21

  • Vertical & horizontal leadership
  • Trust, kindness, willingness, & learning
  • Shared vision & buy-in from whole team
  • Create norms about data & docs, leverage overlapping skills, on/offboarding
  • Join global & create local communities
  • Open source coding
  • Version control
  • Collaboration & distribution
  • Real-time co-development
  • Connection to broader communities

Our path to better science in less time using open data science tools

Lowndes et al. 2017, Nature Ecology & Evolution

How we work

Tooling

tools & practices

People

teams & community

  • Vertical & horizontal leadership
  • Trust, kindness, willingness, & learning
  • Shared vision & buy-in from whole team
  • Create norms about data & docs, leverage overlapping skills, on/offboarding
  • Join global & create local communities

9 of 21

Tooling

tools & practices

People

teams & community

Open data science

This fuels reproducibility, collaboration, & communication...

10 of 21

Reproducibility

  • Motivation: Foremost to increase efficiency day-to-day for ourselves
    • Data preparation/wrangling in R: as critical as modeling
  • Mindset: Always thinking ahead for Future You & Future Us
    • Documentation, onboarding (self and others)
  • Co-design: Workflows & conventions for analysis & reproducibility
    • OHI Toolbox (RStudio+GitHub), SOPs for data mgmt
  • Tip for teams: Tidy data

Our path to better science in less time using open data science tools

Lowndes et al. 2017, Nature Ecology & Evolution

11 of 21

Tidy data has the same structure...

...so you can use existing tools in familiar ways...

...rather than hand-crafting one-time approaches that are hard to iterate, share, or reuse

12 of 21

  • Motivation: Streamline filesharing & documentation, create resilience
    • Make sure everyone has ways to participate confidently
  • Mindset: Analysis is a collaborative effort, not an individual burden or personal craft
  • Co-design: Shared practices for documentation & archiving
    • GitHub & Google Drive (not MS), GitHub Issues (not email)
  • Tip for teams: Regular analysis-centered meetings (“Seaside Chats”)

Collaboration

Our path to better science in less time using open data science tools

Lowndes et al. 2017, Nature Ecology & Evolution

13 of 21

Seaside Chats (regular analysis-centered meetings)

  • Normalize talking about data
  • Identify & address shared needs
  • Think ahead for Future You & Future Us

14 of 21

  • Motivation: Streamline communicating with external audiences. Later: engage stakeholders, efficiently incorporate ideas & feedback
    • Move beyond emailing PDFs to our community
  • Mindset: Reimagine communication as part of our daily workflows, not only upon publication
  • Co-design: The same tools for analysis & communication: R+GitHub
    • Webpages, websites, blogs, e-books, slides, interactive dashboards...
  • Tip for teams: RMarkdown

Communication

Our path to better science in less time using open data science tools

Lowndes et al. 2017, Nature Ecology & Evolution

15 of 21

RMarkdown fuels new open science communication channels with R

RMarkdown files combine code & text

Renders to Word, PDFs, and the web

RMarkdown file renders to:

Word!

PDF!

web!

16 of 21

Tooling

tools & practices

People

teams & community

Open data science

fuels reproducibility, collaboration, & communication

?

17 of 21

Designed to onboard teams to open science through open data science

  • Operationalizing our experiences with the Ocean Health Index
  • Remote-by-design & cohort-based: 3 hrs/month for 2-4 months
  • First cohorts: environmental scientists from academia and NOAA

We empower teams to work in ways that are more efficient, resilient, & inclusive

Part of a larger effort to welcome scientists & grow the open movement

OPENSCAPES CHAMPIONS

An open data science mentorship program for science teams.

18 of 21

“If you want to learn to write, you read a lot,

if you want to play music, you listen a lot.

It’s hard to do this with data analysis.”

- Hilary Parker & Roger Peng

RStudio::conf(2020) keynote & NSSD podcast episode 100

19 of 21

Julia Stewart Lowndes, PhD

http://jules32.github.io

lowndes@nceas.ucsb.edu

@juliesquid

openscapes.org ohi-science.org

@openscapes @ohi-science

  • Environmental assessments
  • Open data science
  • Empowerment, resilience, and inclusion

This is a story about...

Artwork by Allison Horst for Openscapes

Thank You

20 of 21

21 of 21

Shiny enables interactivity for stakeholder engagement & co-creation with R