The holiday busy season has ended, and we are officially back to work. I hope everyone in the northern hemisphere was able to find enough time for cookies and cozy blankets and hygge — and that in the southern hemisphere, you're able to at least appreciate the longer summer nights, despite, you know, climate change.
Please feel free to forward this blog on to anyone you think might be interested.
Let's get to it.
What Even is Data Science, Anyway?
Nothing like starting off a topic that will inevitably have the "well, actually…" horde on my tail in full swing. It’s my own fault for getting cozy after the holidays! OK, but it’s a serious issue anyway.
"Data Science" is going nuts — analytically speaking. Harvard Business Review said "Data Scientist" was the "Sexiest Job of the 21st Century" (but, like, that was in 2012, before we knew about Instagram Face…so let's not hold our breath.)
It's also--I hate to be the bearer of bad news on this one--completely made up. I mean, not like out of whole cloth, but it's also a broad enough umbrella that it includes a LOT of folks (statisticians, data engineers, data analysts, machine learning (ML) aficionados, etc.).
Some have argued that having such a nebulous definition for data science is more harmful than helpful and that organizations would do well to rein in the titles of their positions to reflect the specific specialties that they engage in.
Personally, I kind of like this open-air diffusion of “data scientist” because it doesn't allow as many gatekeepers to dictate that it's just those with a computer science degree who can work in data science. We need to have a more inclusive and diverse cast of characters and skill sets. Similar to, you know, the way that real-life problems require diverse perspectives and backgrounds to actually get solved.
Personally, I really like the description from the Economist that data scientists "combine the skills of software programmer, statistician, and storyteller/artist to extract the nuggets of gold hidden under mountains of data."
Wait — isn't ”data science” old, and now we just talk about AI?
This isn’t a silly question. The concepts are often talked about in similar veins of discussions around "digital transformation.” At Abt, for example, our digital teams often discuss everything from data systems to digital architecture to analytical tools that use various modes of analysis, both traditional in the form of statistical regressions, to today’s thinking about how we can set our data up for deep learning analyses. We have to be fluent talking across disciplines and tools.
I mean at some point we have to be the adults in the room and also address that data science is not in fact a repudiation of statistics. I personally don’t feel like I see this debate as much anymore, but it lingers in the halls of “well, actually…” along with that person on your team who chuckles “you know the cloud is just another person’s computer, right?” Let’s get fewer of these in 2020.
But back to data science vs. AI (and I’m lumping in ML here because we see that everywhere too), which aren't synonymous. In a way they build on each other, as the tools and analysis of a data science methodology provide the required stepping stone for ML tools, which themselves are then joined to develop AI algorithms. There are better and more elaborate elucidations here, here, and here.
Actually, if I were to call attention to anything, it'd be the widening gap between data scientists and data engineers. A lot of folks would love to live in a world where all we have is clean, structured, ready-to-analyze data that we can run analyses on and find connections that change industries. But often the data is messy, fragmented, siloed, and needs to be connected by APIs and interoperability architecture before it even gets to the science-y bit. So -- more data engineer unicorns, please.
So -- which flavor of data professional corresponds to what you're doing/interested in?
Stay tuned – we are working on some material to post that outlines a little more the skills and areas of interest that we are finding valuable on our digital teams at Abt Associates – so this is just the beginning!
Some Other Random Data Science Links
- Data Literacy Is the Key to the Future of Data Visualization — Tableau and beyond.
- Unearthing Lessons by Revisiting—Not Reinventing—the Wheel by Utilizing Data Science — From our colleagues over at DAI – I thought this was a succinct way to tackle a big area – always a fun time!
And of course, random AI links
- Artificial intelligence for Sustainable Development: synthesis report, Mobile Learning Week 2019 — Creepy cover, but a great primer if you feel like you're late to the AI in Development game.
- 2,602 uses of AI for social good, and what we learned from them — There’s a lot to learn here.
- Opinion - AI for Good is Often Bad | Wired — Shots fired! But also -- very much in line with Tech4Good, actually. Technology ain't the answer, but systems are.
- Can a Machine Learn to Write for The New Yorker? | The New Yorker — This one was shared so much with me I feel legally bound to make sure that everyone on this list has seen it.
- Google’s ‘Project Nightingale’ Gathers Personal Health Data on Millions of Americans - WSJ — Cool cool cool.
- Digital technologies and the developing world | FT Tech Tonic on acast — Podcast Episode alert! Economist Stefan Dercon tells John Thornhill about the findings of a research project he led, showing how, used wisely, technology can enable development rather than just replace labor and put people out of work. Neato!
- The Media's Coverage of AI is Bogus - Scientific American Blog Network — AKA why we have this blog.
- Presentation - How to Recognize AI Snake Oil — An antidote to the previous!
- Using AI to Understand What Causes Diseases — Testing your knowledge from the last two!!
- Interpreter, Google’s real-time translator, comes to mobile | TechCrunch — Now in 44 languages!
- Beautiful News — If you need a hit of good data visualization.
- SpaceX Is Sending Super-Muscular Mice to the Space Station — It's better than sending up convertibles, IMO.
Thirsty for more? Check out what else we are doing and talking about when it comes to Digital Transformations.