Following up on the excellent panel at ASIS&T, I’m looking to write a vision statement about what the field of human centered data science could be, viewed from the lens of information science.
Human. Who gets to count as human? Who is seen as the central case of the human cateogry? How can we expand our conception of who humans are? What does counting as human get you in a data science context?
Centered. What other kinds of centered-ness are apparent in data science practice (e.g. money? knowledge?)? How do we re-center on humans?
Data. Data is always partial, both in the sense of incomplete and in the sense of preferential (Brian Cantwell Smith). Data is really capta (Johanna Drucker). How does our disciplinary understanding of what data is reframe data science
Science. Science is constrained constructivism (Katherine Hayles). How can we make better constraints to shape what this science constructs?
I envision this paper as a lens through which to refract great contributions from our literature and reframe them as theoretical underpinnings for what Human Centered Data Science might become.
Another approach, inspired by conversations with Theresa, Passiona, Catherine, Michael and others:
Full stack data science might be a workable ‘brand’ for IS data science programs. Key to this, though, would be defining exactly what layers constitute the stack. The ‘stack’ of data science is usually defined merely technologically, and a strong statement that there are social and ethical layers could be a convincing statement of the IS approach. Best of all, the allusion to a familiar software development role and practice makes this intelligible to industry and interdisciplinary partners without jargon.
A compelling graphic of the stack and its layers could be a succinct summary of the curriculum framework the iSchool data science committee is working on.
A challenge for this term would be that currently ‘full-stack developers’ aren’t necessarily doing human-centered design (e.g., in a web project, they wouldn’t do UX research, discovery or design). ‘Full-stack’ seems to denote merely technical skills, albeit a wide spectrum (front+backend).
That is absolutely the challenge. The world doesn’t speak our language. For example, even ‘sociotechical’ isn’t really meaningful for most.
So we’ve got to get into the conversation about what these terms should mean, and share our positive vision for them. I’m not saying ‘full stack data science’ is 100% one of these terms, but I think it might be. The notion of ‘full stack’ currently means a set of technical skills corresponding to technologies. Our contribution is that this stack isn’t ‘full’ unless it includes the social, ethical, and political layers that we’re good at.
This ultimately connects the education discussion to broader social and political concerns about technology, and sets a vision for how our disciplinary contributions build this bridge.
The problem with everything I said above is that it took several paragraphs. We need a snappy diagram that says all of this with a picture.