Brendan Le Grange

Personal projects are a great low-stakes way to touch all of the other points mentioned here: they give you a place to practice the skills you're learning on your online course, a reason to join and participate in new communities, a hook with which to approach mentors in other arenas, and, if you do something really interesting, a way to get invited to talk at one of those conferences. I'm starting a new podcast in the AI space and it has been a tremendous motivator to get me reading books and testing tools I'd been putting aside for far too long!

I've just recently seen Guy Kawasaki's 10-20-30 rules doing the rounds again: no deck should have more than 10 slides, take more than 20 minutes to present, or use a font smaller than 30pt. The first two translate pretty directly and while font size doesn't really come into visualizations, the principle still holds: make the image you choose incredibly easy to see. Use fewer and bolder lines, ensure the scale is understandable and directionally logical, and if you need a lot of technical data put that in a handout instead.

Editing is vital, strip your story down to only its strongest lines of logic. It can be tempting to add extra support because surely one and a half reasons are better than just one, but as soon as you present an argument you invite your audience to think about reasons why it could be wrong and if they find one in their minds, even if it is just for one out of three or four or five reasons presented, they'll either waste a lot of time asking about it or, worse, they'll keep quiet but lose faith in your entire case.

Even before data collection comess the question of what data to collect. One big and obvious benefit of a diverse team is that it brings a diverse set of originating questions and fewer shared invisible assumptions.

An old trick I still use is to write the headings of every slide as the concept it is trying to communicate: instead of 'new credit card volumes' write 'we issued a record number of credit cards' - then you should be able to read only the headings of each slide and understand the presentation message: 'we issued a record number of credit cards', 'most of the growth came from millennials opening gold cards' 'the risk profile is inline with historic norms' 'as are early vintage defaults' 'though average limits are 10% smaller' 'as such I suggest testing an early credit limit increase program'

Don't create a presentation that shows all the work you did, create a presentation that shows what that work taught you. It is an inescapable fact that half of your work is foundational, or worse, a dead end, that's OK but your manager should trust you enough to not need to walk those deadends behind you - if you did weeks of work and run a dozen bits of analysis to finally understand one chart's worth of concept, just show that one chart.