The Berkeley Institute for Data Science: a place for people like us

Video of presentation: https://www.youtube.com/watch?v=q5yAy4WWTyU

BIDS was created as a novel environment for Data Science, in collaboration with the University of Washington in Seattle and NYU, with funding from the Gordon and Betty Moore Foundation and the Alfred P. Sloan Foundation.

Transcript

1. BIDS: Berkeley Institute for
   Data Science
   “a place for people like us”
   Fernando Perez
   @fperez_org
   

   View Slide

2. “People like us”?
   (a careers problem in academia)
   • Folks at the intersection of domain science, methods and
   software.
   • We actually care about our code
   – Not just about our results/publication count.
   • We often belong to the Department of Connective Tissue
   – Good luck getting tenure in that one!
   • The Big Data Brain Drain: Why Science is in Trouble.
   – Great blog post by Jake van der Plas.
   

   View Slide

3. A ‘home’ for people like us?
   Credit: @ananelson of dexy fame.
   

   View Slide

4. Raise your hand if…
   Your dept. chair/dean/provost
   • Loves that you write lots of code and therefore less
   papers.
   • Loves that your papers have many authors from
   different departments and you’re lost in the middle.
   • Loves that the journals you’ve published in seems like
   a random collection of unrelated topics.
   • Encourages you to spend more time on mailing
   lists/github/hipchat/IRC helping random strangers.
   • …
   

   View Slide

5. Catalyst: Moore/Sloan Initiative
   • 3 chosen out of 15
   LOIs invited
   • 5 Yrs, $38MM
   • Announced at White
   House OSTP event
   Nov 2013
   

   View Slide

6. U Washington eScience: SciPy’14 Sponsors!
   • Sarah Stone
   • Bill Howe
   • Jake van der Plas
   

   View Slide

7. NYU Center for Data Science: cds.nyu.edu
   

   View Slide

8. •Moore/Sloan Initiative: Core goals
   •Support meaningful collaborations between
   •Methodology fields: Comp Sci, Stats, Applied Math
   •Science domains
   •Establish sustainable Data Science career paths
   •A new generation of multi-disciplinary scientists
   •A new generation of data scientists focused on tool development
   •Build an ecosystem of tools and research practices
   •Sustainable, reusable, extensible
   •Effective as scientific research tools
   8
   

   View Slide

9. Initial Data Science Faculty Group
   Joshua Bloom, Professor, Astronomy;
   Director, Center for Time Domain
   Informatics
   Henry Brady, Dean, Goldman School of
   Public Policy
   Cathryn Carson, Associate Dean,
   Social Sciences; Acting Director of Social
   Sciences Data Laboratory "D-Lab”
   David Culler, Chair, EECS
   Michael Franklin, Professor; EECS, Co-
   Director, AMP Lab
   Erik Mitchell, Associate University
   Librarian
   •Faculty Lead/PI: Saul Perlmutter, Physics, Berkeley Center for Cosmological Physics
   •Fernando Perez, Researcher, Henry H.
   Wheeler Jr. Brain Imaging Center
   •Jasjeet Sekhon, Professor, Political
   Science and Statistics; Center for Causal
   Inference and Program Evaluation
   •
   Jamie Sethian, Professor, Mathematics
   •Kimmen Sjölander, Professor,
   Bioengineering, Plant and Microbial
   Biology
   •Philip Stark, Chair, Statistics
   •
   Ion Stoica, Professor, EECS; Co-Director,
   AMP Lab
   

   View Slide

10. Berkeley Institute for Data
    Science (BIDS)
    Relevance across the campus
    suggests need for central
    location that will serve as home
    for data science efforts
    Doe Library
    Enhancing strengths of
    • Simons Institute for the
    Theory of Computing
    • D-Lab (Barrows)
    • AMP Lab (EECS)
    • CITRIS
    • SDAV Institute (LBL)
    • Urban Analytics Lab
    • etc.
    

    View Slide

11. Doe Memorial Library
    Doe Memorial Library at
    the heart of the UC
    Berkeley campus will be
    the new home of the
    Berkeley Institute for Data
    Science (BIDS).
    The campus has set aside
    5,000 sq ft on the ground
    floor directly accessible
    from the building’s north
    entrance and opposite to
    the historical Morrison
    Reading Room.
    

    View Slide

12. Exec. Director and Data Science Fellows
    Exec. Director: Kevin Koy
    • Berkeley Geospatial Innovation Facility (GIF)
    • Numpy, Pandas, Django tools and REST APIs for GIS in research.
    Fellows affiliated with this community (15 total):
    • Katy Huff (SciPy)
    • Dav Clark (presented Tuesday)
    • Kyle Barbari (AstroPy)
    • Karthik Ram (Software Carpentry, ROpenSci)
    • Justin Kitzes (Software Carpentry)
    

    View Slide

13. 13
    Applied Math
    /
    Working Groups as Bridges
    The collaboration model across NYU-Berkeley-UW
    • Software Tools and
    Environments
    • Reproducible Research
    and Open Science
    • Education and Training
    • Ethnography and
    Evaluation
    • Career Paths
    • Space and Culture
    

    View Slide

14. Software Tools and Environments
    • Tools for open, reproducible data science:
    – Python, Julia, R, Scala, etc.
    • Our three universities:
    – Deep expertise we don’t always engage.
    – Space and resources.
    – Complement github/OSS models.
    •14
    

    View Slide

15. Reproducible Research and Open
    Science
    •15
    • Incentive models
    –funding, publication
    • Build tools and practices.
    • Ask the right epistemological questions
    (cf. Lorena)
    

    View Slide

16. Education
    •16
    • What are the conceptual foundations of
    data science?
    • How do we teach them?
    • Are there new programs
    (undergrad/grad) or just courses?
    

    View Slide

17. Career Paths
    •17
    • Stable, rewarding, competitive.
    • New criteria for faculty
    tenure/promotion & scholarship?
    • New kinds of faculty FTEs?
    • New kinds of Research Software
    Engineer positions?
    

    View Slide

18. Ethnography and Evaluation
    •18
    • Study the process itself of doing data
    science.
    • This project seeks institutional change.
    Talk to Dav Clark and Seb Benthall.
    

    View Slide

19. Industry Outreach
    • Problems that cross the industry/academia lines
    • Data/resources are often in industry
    – Though see Facebook’s recent experience…
    • Not enough academic jobs:
    – Many of our students will go to industry
    – Good collaboration and recruitment opportunities.
    •19
    

    View Slide

20. Questions for discussion
    • Is “Data Science” actually “a thing” in science?
    • Is something like BIDS the right intellectual &
    institutional space for many folks from the
    SciPy community?
    • How can we (BIDS) better engage and support
    the SciPy community?
    

    View Slide

21. For more context, see this blog post
    An ambitious experiment in Data
    Science takes off: a biased, Open
    Source view from Berkeley
    

    View Slide