Skip to end of metadata
Go to start of metadata

Aaron Culich (AmpLAB) and Dav Clark (D-Lab) attended a workshop on Nov 14-15, 2013 that recapped the accomplishments of the DSSG Summer Institute in Summer 2013. They report-back to the reading group on Th 21 November.


Four one-page case-studies (see attached PDFs: case.pdflandbank.pdfjail.pdfcta.pdf)

Optional Readings:

list of projects on the DSSG web site

Github repo and wiki for the "Growth Curves" DSSG project


Questions to consider:

  • Are there centers or organizations on the Berkeley campus interested in engaging in similar projects with a focus on the public service aspect of UC's mission?
  • What IT resources would facilitate such an engagement, and how might they be most effectively and efficiently provided?

ATTENDING: Dav Clark, Chris Hoffman, George Atala, Rick Jaffe, David Greenbaum, Aron Roberts, Sharon Goetz, Cathryn Carson, Patrick Schmitz, Aaron Culich, John Lowe, Richard Millet, Steve Masover


DSSG identified project partners & research fellows first; problem identification might be a place to start.

Involved policy and journalism people communicating with partners (e.g., gov't agencies with questions to answer/problems to solve), and communicating out results of the summer project.

Dav: I like to see projects that continue ... but concern that we end up with a set of projects for which we had grant-supported infrastructure, but then at end of grant they're left hanging. What's an appropriate sustainability strategy.

Dav: I will likely take a lead on determining whether UCB is interested in doing a DSSG like set of projects, and how.

Aaron: Worth noting that people who ran last summer's institute were three very different people there for very different reasons/agendas. Went from something quite chaotic to something focused and organized; interesting to see how the effort self-organized.

Dav: Broadly speaking:

  1. primary lead (Rayid Ghani) was also primary organizer of KDD @ ACM. Was skeptical of utility of political action; interesting in context that a lot of the energy for this effort came out of the Obama campaign-- a visionary.
  2. Matthew Gee: another Obama campaign person, junior in academic career but very enthusiastic about political action -- mediator between Raid's visionary dispostion and Juan's get-it-done attitude.
  3. Juan-Pablo Velez -- studied journalism, a coder / auto-didact, really disposed to get his hands dirty -- had a lot to do with putting the wiki together and organizing it -- a closer.

Patrick: Why wouldn't CITRIS be the right org on the Berkeley campus for this space?

Dav: We're here only because this is the first group we've reported-back to, no decision made about who's involved, where the right home is – presence here doesn't preclude other groups involvement/lead. Berkeley Inst of Design, et al. as well as CITRIS.

Aaron: A key element here is to determine what are a good set of partnerships to develop, to fuel and focus this effort, and how do we achieve those partnerships.

Patrick: Code for America?

Dav: Juan is involved in that. Allied, different model.

Aaron: "Danger zone" in intersection between hackers & gov't? (Drew Conway DS venn diagram)

Patrick/Rick/David: All kinds of partnership already seeded between UCB campus and city/gov't/public

Dav: Code for America is very general; DSSG is focused in testing assumptions about what Data Science can do in useful, practical terms.

Aaron: DSSG stayed small, deliberately. 50 people. One office space. A cohort model, which was important, something valuable in that effort. Selective invitation to this workshop we attended, people who the DSSG folks thought might replicate this model successfully and usefully.

Patrick: Hack the City in SF. Not sure who here does that, what happens here.

Cathryn: Paul Waddell, Urban Analytics.

Aaron: The best big data problems are those in which the partners are engaged, not bystanders or tourists, not just riding along.

Patrick: So start from an interesting question, and see if you can come up with tools to answer that question

Dav: We should be looking for problems where the data available to analyze is clearly defined and available.

Aaron: DSSG said they were interested in solving real problems; but their MISSION was education and training of the cohort, the 50 people who participated. That they happened to solve a real problem along the way: bonus, kudos.

Aaron: Several flavors of sustainability. Organizational (mentors, fellows, partner organizations); IT (code, data, maintaining an application that grew out of a project scoped to a summer).

Dav: Federal gov't can be a difficult partner because they dither and hedge, sometimes in response to politics of the moment ...

Aaron: public relations issues, legal issues -- can be torpedoed by elements of an organization one doesn't even foresee, necessarily.

George: End product?

Dav: Meant to be academic, proto-academic paper. I'd say there were three aspects of outcome: education/training for the participants; outcome relevant to the partner -- the answer to a question; communication of outcome.

Aaron: Product was the insight gained as a result of the analysis, embodied in the report produced as a specific deliverable from the project team.

Aaron: At first summer institute, participants felt no need to be efficient w/ AWS instances because they had "free" compute, a grant. Next time, could be more thoughtful and efficient about use of cloud resources.

Patrick: Reuse of project outcomes? E.g., many cities have bike share programs.

Dav: The DSSG folks didn't feel they were equipped, funded, resourced to figure out how Divvy (bike share project) could be applied elsewhere.

Aaron: Outreach expressed, perhaps, in DSSG reaching out to Berkeley, Columbia -- places that might take this type of effort outward and forward in time.

Cathryn: Domain expertise. Berkeley would want to bring that on-board before proceeding.

Dav: Let's use this kind of effort to crystalize the incohate set of talents that we have on the Berkeley campus. At this point it's about defining a clear mission that can catalyze that crystalization.

David: Organizations in Oakland (non-profits) who might be good domain-centered, urban focused partnerships.

Dav: Have been talking with Jon Stiles about bringing City of Richmond in too, for all the obvious, good reasons, plenty of opportunity there.

(Many): next step is to seek a cohort of supporters/stakeholders/partners who are interested in making some public-service focused set of projects to which Data Science methods and resources can be applied.








  • No labels