Human Learning

At Scale


Kathryn (Katy) Huff

Chicago, IL

August 27, 2016

cyclus pyrk pyne THW Software Carpentry Lab Carpentry Logo book ARFC Logo BIDS Logo

The world is data-intensive and computational.

Humans are clamoring to learn how to live in it.

hpc results

The Hacker Within is Helping

Peer-led organization for scientific computing best practices.

THW at BIDS
Nuclear Engineering and Engineering Physics, University of Wisconsin - Madison Nuclear Engineering, University of California, Berkeley melbourne Michigan State Yale swinburne UIUC THW

Community

  • democratic anarchy
  • maximize permissions
  • charismatic leadership
  • occasional beer
excellent team

Peer-Driven

  • peer teaching and learning
  • interdisciplinary normalizes mixed skill levels
  • relevant topics
  • appeals to the desire to show off
  • direct peer-to-peer skill sharing
SC Bootcamp 2
SC Bootcamp 1
Python Bootcamp
hunter
scbc

Software Carpentry is Helping

Software Carpentry

Software Carpentry Foundation

  • NumFOCUS umbrella
  • Elected Steering Committee
  • Representative, Expert Advisory Council

We make researchers in science, engineering, and medicine more productive by teaching them basic lab skills for scientific computing.

workshop

Software Carpentry Workshops

  • Two days of hands-on learning
  • Scientists teaching scientists
  • Instructors are volunteers
  • Materials are all open access
  • bash --> automate tasks
  • python --> build modular code
  • git --> track and share work
  • SQL --> manage data
  • nose --> program defensively
workshops

But Why Do They Do It?

  • community
  • travel
  • teaching experience
  • teaching = learning
  • sensed need
  • etc.
workshops
instructors vs workshops
workshop sizes

New Resources Are Helping

book

Getting Started



  • Chapter 1 - Introduction to the Command Line
  • Chapter 2 - Programming Blast Off with Python
  • Chapter 3 - Essential Containers
  • Chapter 4 - Flow Control and Logic
  • Chapter 5 - Operating with Functions
  • Chapter 6 - Classes and Objects




Getting It Done


  • Chapter 7 - Analysis and Visualization
  • Chapter 8 - Regular Expressions
  • Chapter 9 - NumPy: Thinking in Arrays
  • Chapter 10 - Storing Data: Files and HDF5
  • Chapter 11 - Important Data Structures in Physics
  • Chapter 12 - Performing in Parallel
  • Chapter 13 - Deploying Software


Getting It Right



  • Chapter 14 - Building Software Pipelines
  • Chapter 15 - Local Version Control
  • Chapter 16 - Remote Version Control
  • Chapter 17 - Debugging
  • Chapter 18 - Testing


Getting It Out There



  • Chapter 19 - Documentation
  • Chapter 20 - Publication
  • Chapter 21 - Collaboration
  • Chapter 22 - Licenses, Ownership, and Copyright
  • Chapter 23 - Further Musings


Effective Computation Course

  • Open source IPython Notebooks
  • Project-Driven Course Structure
  • Example-Driven In Class Work
  • BYO-Science Template
  • Alpha-testing, Tiffany Timbers, Simon Fraser
  • Beta release soon

github.com/physics-codes/seminar

New Curriculum is Appearing

Data Science 8

databears

Cathryn Carson, David Culler, etc. databears.berkeley.edu

Data Science MOOCs

data science mooc at illinois

Rob Rutenbar, John Hart, etc. cs.illinois.edu

But it's not enough

We all have to teach.

All the time.

Lower the Barrier, Not Standards

“The feedback giver explained that he 'used high standards' in evaluating the essays for publication in the teaching magazine. Still, he said, having read the student’s essay, he believed the student could meet those standards. His criticism, this form of feedback implies, was offered to help the student meet the publication’s high standards. Black students trusted this feedback as much as white students, and trusting it powerfully motivated them to improve their essay.”

-- Claude Steele, Whistling Vivaldi

Easy Suggestions

  • New Developer Instructions
  • Comprehensive Documentation
  • Accessible Listhost
  • Define Expectations Clearly
  • Curate Low Hanging Fruit
  • Give Permissions Generously
  • Strict Issue Assignment Workflow

Harder Suggestions

  • Targeted Sprints
  • Appoint an Ambassador
  • Mentor Users Directly
  • Consider Users Conferences

Acknowledgements

Software Carpentry THW

This book was heavily inspired by the work of Software Carpentry and The Hacker Within. We are thrilled to see Effective Computation in Physics come to press and spread the joys of computation further into the physical sciences. In case you were wondering, the cuddly creature on the cover is a bobtail squid.

Acknowledgements

  • Anthony M. Scopatz
  • Paul P.H. Wilson
  • Gregory V. Wilson
  • Lorena Barba
  • Tiffany Timbers

THE END

Katy Huff

katyhuff.github.io/2016-08-27-pydata
Creative Commons License
Human Learning at Scale by Kathryn Huff is licensed under a Creative Commons Attribution 4.0 International License.
Based on a work at http://katyhuff.github.io/2016-08-27-pydata.