On data careers
Resources for newcomers to data science roles
This month's edition is one that takes a short detour from programming tools for data scientists and ventures into career topics. It's something that's been on my mind for a while as I pondered my own career development.
Advice for newcomers looking to break into data science
Firstly, I wanted to highlight recent writing by Lawrence Gray, a data scientist whom I met at PyData NYC 2017. Or was it 2019? It all feels like eons ago by now! From his own unique voice and perspective, Larry has started a collection of new blog posts with advice on building out a data science career. What I like about his viewpoint is that he approaches a data career like an artisan would - centering it around a portfolio of digital artifacts and opportunities that demonstrate your skillsets. This stands in contrast to an approach that focuses on certifications and courses. Data science is an applied and practical discipline, and while courses are useful for picking up skills, it is in applied projects where the nuance, judgment, and decision-making really shine. Go follow Larry on Medium or LinkedIn for insightful advice!
Life of a data scientist
My ex-colleague from NIBR, Anna Kostikova, launched a YouTube channel! Apart from her video editing style being totally hilarious, she also put out lots of valuable advice for newcomer data scientists. Here's one video, 5 tips for interviewing for DS jobs, which I thought was a wonderful watch. Anna is highly experienced and has lots to offer, so keep, ahem, watching her space!
Thinking strategies for data scientists
Secondly, I wanted to highlight a very well-outlined Medium post titled, "5 types of thinking for a high performing data scientist". Though the post is light on details, the outline is fairly comprehensive. Indeed, to deliver impact as a data scientist (defined more closely to the definition of a scientist, and not an engineer), one needs to possess different modes of thinking to be effective. Systems thinking is one that I wanted to highlight, as it is one that I noticed as being a very rare trait. Systems thinking demands looking at interconnections between subsystems and entities and their dynamics, rather than the i.i.d. static snapshots that we might be accustomed to in statistical learning. Sometimes taking that broader view leads us to simpler and more effective solutions. (The opposite can be true too.) Definitely check out the post to learn more.
Software design for data scientists
Greg Wilson (of the Software Carpentries fame) has put out a very cool talk titled "Software Design for Data Scientists". I would encourage you to look at the slides for inspiration. The reason this talk is important is that the longer I stay in the data science world, the more I come to the recognition that a lot of the value proposition of modeling is quite similar to the value proposition of software design - automation - and the way that modeling work delivers a sustainable and lasting impact is very similar to software too. We are, in effect, building systems that automate the delivery of actionable insights, just as software automates the delivery of information. So learning software skills absolutely matters. Greg's talk doesn't drop names on tools but rather outlines key ideas and principles from the software development world that he thinks should translate over to the data science world. Slides are worth flipping through!
From my collection
Two things to highlight from my own collection:
Firstly, the team I'm on is hiring! I wrote a blog post that links out to the relevant posting. In that post, I also took the time to write an FAQ that I thought might be handy.
Secondly, just in July, I release a new version of
nxviz, a graph visualization library, and I am very excited to share the release. Its API design is much more principled than before! Come read the blog post!