Data science venues are weird. Not bad, just awkwardly positioned between machine learning theory and domain applications, and the conferences reflect it.
Tier A in CORE terms means selective, well-cited, respected, one step below the A* flagships. For most researchers, most of the time, tier A is the sweet spot. Real prestige, better acceptance odds, and reviewers who actually read the paper instead of triaging it in eleven minutes. I'll take that trade most years.
The timeline spread is the trick here. Unlike ML, where deadlines cluster, data science conferences scatter across the year. That's good news: you can almost always find a strong venue with a deadline that fits your results, instead of contorting your work to fit a deadline. The catch is that scattered conference submission deadlines are exactly the ones that slip. Run them through a lab deadline dashboard SaaS if you're juggling student projects, or at minimum a venue matcher SaaS that pairs your topic with venues and their dates in one view.
Watch the scope drift. Data science venues have been absorbing everything from deep learning to visualization to fairness research. Read recent proceedings carefully, because what a venue published three years ago may not match what it wants now. Venues evolve faster than their reputations do. I keep saying this to students and they keep not believing me until the desk reject arrives.
And do the boring logistical check. Location, registration cost, whether the conference is in a city your department will fund. A perfect venue you can't afford to attend isn't a perfect venue.
The submission-date calendar at confjournal.fyi lists tier A conferences in data science with their dates, so you can plan your writing backwards from a real deadline.
When the paper is ready, confjournal.fyi will get it submission-ready with paper formatting and submission checks for your target venue.