Last week I had the pleasure of speaking at the annual conference that Harvard’s Center for Geographic Analysis holds. This year the theme was Space & Time in Data Science, and panelists shared stories and nuggets of wisdom for the audience of geographers, geographic information scientists, computer scientists, statisticians, data scientists, and others. Upon prompting for a show of hands about who fell into the different disciplinary categories, many confessed to wearing multiple hats among those roles. Which I think was one point of this event: to foster multi-disciplinary conversations in a place where there aren’t enough going on naturally.
Some of the more noteworthy comments were from:
- Francisca Dominici, a biostatistician and co-Director of Harvard’s Data Science Initiative, whilst talking about methods for causal inference and scientific reproducibility, wondering whether in fact there exists *anything* that we can really control so we can make inferences about today’s world. She described the CGA as an entity able to help connect the data science talents across campus.
- Peter Fox, from RPI. He shared the success that the knowledge network behind the Deep Carbon Observatory has been and was refreshingly forthcoming in his description of how attempts at a University Network of Things hasn’t worked. I am increasingly interested in research infrastructure, and knowledge networks are an important component. As an aside, they have a GIS for Science class at RPI but nothing from the syllabus distinguishes it from basic intro GIS course that uses open-source software and apps.
- Amelia McNamara who had a fountain of ideas I liked, including the notion of an “interactive essay” – like this one one Exploring Histograms. I will definitely be having my students play with this Spatial-Aggregation Explorer. How Spatial Polygons Shape Our World (YouTube link) officially makes her an honorary geographer in my book. Except I’m not sure she wants to be one. She’s doing just fine with her own disciplines.
I had the second-to-the last slot in the last panel of the day. My own comments focused on the role of strategic communication for strategic bridge building (to better connect GIScience & data scientists). Strategic was to be the key word. I’d say four of my five ideas were reasonably on target but one went up in flames rather spectacularly.
I happen to know one (very bright, very engaged) data scientist who works at a data science company in the Silicon Valley, one that I’d never heard of before (or until recently, since). During a conversation with him earlier this year, I learned that he doesn’t know anything about GIScience AND he’d be interested in knowing more. That was that, and I totally forgot the name of his company until I looked him up again while preparing my talk.
So, on Friday afternoon I said that “data science start-ups might be a good place to broker some worthwhile conversations about GIScience,” and I included a screenshot from the website of the company I’d been holding up as an example, vis a vis their young data scientist who expressed curiosity about GIScience: Palantir.
It was late on a Friday afternoon, at the very end of a long day of intellectual prompts, technical rigor, and gobbledy-gook jargon. Brains were noticeably over-saturated. Time remaining only for a few questions or comments for the panelists. The first person who spoke is a GIScientist known for her critical (i.e., in the academic sense) observations. At that moment I really had no idea what she was saying. Her language may have seemed extra circuitous because my brain was tired or she was politely trying to be less direct. The only thing I really heard was her final emphatic statement that “… we’re not going to work with Palantir!”
Wait, what? She knows the company too? Yup, that Palantir. That’s the one. The one that I suggested to a crowd at Harvard that we GIScientists ought to play more with in the sandbox. Maybe not so strategic after all.
I was nicely wisened up by a few folks as we were departing the conference. In the big scheme of things, as we say in Portuguese, não faz mal.
But I’m left with a bunch of conflicted feelings. I still think that conversations with data scientists at start-ups are a good thing. Not everyone working at Company P is mal-intentioned and sneaky, especially and definitely not my data scientist friend. Life is what you do, not what you say, so we let our actions speak for themselves. I spend way too much time sitting in a small home office by myself in a centrally-isolated patch of land in upstate New York. I crave the chance to develop and brainstorm ideas for talks with colleagues and within a community of practice. I sometimes learn from my mistakes.