imho the Coding Train videos are well prepped but the production aspect and editing is minimal. for theory-heavy topics, having a clear throughline is essential for me, even if the presenter does stumble every now and then with specifics

im getting a 404 on the link

might be a stretch but paging @attilabatorfy.bsky.social as the first (second?) of the six degrees of separation

i feel uneasy saying this, but they are visually spectacular

so, this is the final form?

I'd argue it's truly minimal, we even managed to remove any morsel of meaning from it

I 'Um, Actually"-ed so hard, I got the math wrong... oh well

if you're gonna argue with minimalism, which is simply The Right Way, you should do it without standing up a straw man.

the optimal solution here is clearly using Braille for the column labels, which reduces the true data-ink ratio from a measly 1:45 (yours) to an impressive 1:24 (mine). </sarcasm>

where did you move to? 😮

imho the issue is more that data professionals sit in central teams, instead of working closely with the people our craft is ought to make more productive.

which, in turn, is understandable because we are a big overhead cost, that is - as you note- difficult to calculate returns on

i think the "commodification" of dashboards also plays a role in this devaluation. creating a new dashboard is becoming cheaper and cheaper in orgs, so their average value to users is also going down.

for our most recent project, our designer sent over a css file with all the colors as properties (11 tones each of about 7 hues). I worked a bit with them to clean up the system, basically bring it closer to how tailwind does their colors and finally converted them to SCSS vars

I'd say always the one you're most comfortable with! unless you want to learn while doing it, in which case the one that interests you most.

i don't but i also think the changes have become a lot less predictable over the last couple years. many old heuristics like "it's cheapest x weeks before" don't seem true anymore.

I made this one over a decade ago now

www.darkhorseanalytics.com/blog/data-lo...

I know the answer is prolly quite involved but... how does it work? Is it deterministic?