How the Ocean Health Index enables better science in less time

Julia Stewart Lowndes, PhD

Marine Data Scientist & Mozilla Fellow

National Center for Ecological Analysis

& Synthesis, UC Santa Barbara

slides: jules32.github.io

twitter: @juliesquid

April 15, 2019

Bren School Seminar

UC Santa Barbara

Allison Horst

Public

Allison Horst

Allison Horst

3

Public

Allison Horst

“Data science is the discipline of turning raw data into understanding”

Hadley Wickham

Statistician, Professor, Developer

Chief Scientist, RStudio �

“Open data science involves mindsets and skillsets emphasizing efficiency, reproducibility, transparency, collaboration, communication, and kindness”

Me

Outline

oceanhealthindex.org

Halpern et al. 2012

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

​

​

• Captures coupled system health, incorporates sustainability
• Boils into easy-to-understand metrics
• Is flexible to different contexts
• Stimulates actions to improve ocean health
• Is repeatable to track progress through time

Halpern et al. 2012

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

​

Important because in marine management there is need for:

​

• Science- and data-driven methods to measure what people care about
• Standardized but flexible methods to assess different geographies
• Streamlined assessments from year-to-year to track change through time

Assessments we lead

Assessments we enable

OHI+

A healthy ocean sustainably delivers a range of benefits to people now and in the future.

Employment

Cultural Identity and Sense of Place

Food Provision

goals

OHI framework

A healthy ocean sustainably delivers a range of benefits to people now and in the future.

goals

scores

models

data

inform

Repeatable OHI assessment process

​

​

Combining datasets

Documenting decisions & methods

Asking for & incorporating feedback

Collaborating

Revising and comparing

Analyzing and summarizing data

Reevaluating past decisions

Critically evaluating results

Communicating methods and results

Planning

Designing figures

Gathering data

Reading the literature

OHI, aka modern science

ohi-science.org

@ohiscience

vs.

data_final_final.xls

​

Re:FWD: data question

scripts/species_count.R

​

Issue: species count

​

Allison Horst

How we work

​

​

• Culture of sharing, teaching, learning
• Shared practices & norms
• Seaside Chats
• Resilience
• On/off boarding

connection to broader communities

• listen
• learn
• contribute

​

open coding language

• shared practices
• streamlined tools

​

collaboration platform

• bookkeeping
• display, publication, & distribution

​

docs, slides, sheets

• coordinate
• co-develop
• share

Streamlined workflow

Figure adapted from Teucher 2018

Shared tools & practices

efficiency & reproducibility: coding and version control are the keystone

Shared tools & practices

but also for

game-changing

collaboration & communication

efficiency & reproducibility: coding and version control are the keystone

Shared practices for reproducibility

Wickham & Grolemund 2016

RStudio for R, text editing, Github sync, and more

Shared coding practices; convenient interface for coding and syncing

R code (scripts and console)

File navigation, help,

plots, packages

GitHub connection, environment, build

Github for archiving & bookkeeping

​

​

Convenient sharing with yourself and others

See what changed line-by-line

...and plot-by-plot

Github for discussion & project management

​

​

Individual

to do’s

Shared & archived

conversations

Project management and institutional memory

R + Github for documentation & communication

Website

OHI-Science.org

Protocols & methods

*made with RMarkdown*

Hands-on Training Books

R + Github for publication

Interactive

applications

*made with shiny*

ohi-science.org/ohi-global

Shared tools & practices

Our workflow is more streamlined; efficient onboarding & offboarding

Enabling better science in less time

• Culture of sharing, teaching, learning
• Shared practices & norms
• Focus on training & building community

Learning with online communities

https://blog.mozilla.org

openscapes.org

@openscapes

Empower

Amplify

Engage

Allison Horst

We champion open practices to help uncover data-driven solutions faster.

Build champions and communities

Build confidence and skills

Build awareness and excitement

Openscapes Champions

​

We help Champions & their labs:

​

• Engage through culture of sharing, teaching, learning

​

• Be empowered through guidance, pathways & agency for skill-building

​

• Amplify their efforts

Mentorship program that empowers environmental scientists

with open data science tools and grows the community of practice

Openscapes Champions

Lessons based from Lowndes et al. 2017

Openscapes.org/series

Early lesson:

• Data science as a discipline
• Open data science tools exist
• Open as a way to work
• Lab members as a team
• Collaborators and community (redefined) as a way to learn
• The internet as an underleveraged tool for science

Openscapes

​

• Half-way through inaugural cohort with seven labs!
• Seaside, Bluffside, Bayside, Fishbowl Chats
• Openscapes.org resources, blogs, series
• Champions Summit!
• Long-term plan – annual concurrent cohorts, in-person workshops

Progress and what’s next:

What can you do to engage with open data science?

Allison Horst

What can you do to engage with open data science?

- Talk about your data challenges with colleagues

- Share your next presentation online

- Use Twitter for science

- Follow selectively, listen &

learn (e.g. #rstats, @nceas)

​

​

​

1. Promote/enable the culture of open data science – even if you don’t code

​

2. Create/join communities, locally & online

​

3. Use existing online resources to learn & skillshare

• ohi-science.org/betterscienceinlesstime
• openscapes.org/resources

​

4. Ask for open data science skills to be formally taught

How about TODAY:

So what can you do?

A few of UCSB’s many learning communities

eco-data-science.github.io

@ecodatasci

meetup.com/

rladies-santa-barbara

@RLadiesSB

#TidyTuesday Hacky Hours

ESM 206 & 244

library.ucsb.edu/

software-carpentry

NCEAS opportunities: internships, postdocs, research scientists:

nceas.ucsb.edu/employment

We can get to environmental solutions faster

​

if we are more

efficient & collaborative

with how we do science.

​

​

Let’s do

better science

in less time together.

​

​

​

​

Julia Stewart Lowndes, PhD

http://jules32.github.io

lowndes@nceas.ucsb.edu

@juliesquid

​

openscapes.org ohi-science.org

@openscapes @ohi-science

slides: jules32.github.io

Thank You

Extra slides

Using twitter for science

​

My internal monologue:

​

1. Cool visualization!
2. I want to represent my data this way
3. He includes his code!
4. Package from @sckottie at rOpenSci
5. rnoaa is a package making NOAA data more accessible!

​

Using twitter for science

​

Being open with your science

• Learning & getting feedback
• Connecting with people
• Less reinventing the wheel
• Open mindset: expecting better ways and sharing your approaches

Twitter to learn and connect

• Follow selectively & deliberately, listen & learn
• Learn R (e.g. #rstats)
• Follow colleagues. Tell someone you like their paper.
• Find fellowships & jobs.

Shared practices for reproducibility

Wickham & Grolemund 2016

Data wrangling: up to 50–80% of a data scientist’s time Lohr 2014

Tidy data

What if you needed D. opalescens 2017?

Untidy :(

Tidy !!

Good for data entry, not good for data analysis because:

• Data are in column headers
• What are the values?
• Variable is spread across multiple columns

Great for data analysis because

• Each variable has its own column.
• Each observation has its own row.
• Each value has its own cell.

​

​

Tidy data

Examples from tidyr

tidyr::gather()

separate()

gather()

Our Ocean Health Index story

Ocean management is complex

​

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

​