"Is DSA needed for data analyst roles?" is one of the most-searched and worst-answered questions in placement prep — because the honest answer is "it depends which data role," and most advice flattens three very different jobs into one. A data analyst, a data scientist, and an ML engineer face wildly different coding bars. Here's the breakdown, so you don't grind graph algorithms for a job that mostly wants SQL.

Short version: the more "engineering" in the title, the more DSA. Pure analyst → mostly SQL and stats. ML or Data Engineer → real DSA, close to an SDE loop.

Data Analyst — SQL and stats outrank DSA

For most data analyst roles the interview centre of gravity is SQL, spreadsheets and pandas, statistics, and case-style "how would you measure X" questions — not LeetCode. You should still write clean, correct code, but heavy data-structure problems are uncommon. Time spent mastering window functions, joins, and group-bys pays off far more than grinding trees.

• Go deep: SQL (joins, aggregations, window functions, subqueries), pandas, descriptive statistics, and A/B-test reasoning.
• Light DSA: arrays, strings, hash maps, and basic complexity — enough to clear a simple coding screen.
• Rarely: trees, graphs, dynamic programming. Don't lead with these.

Data Scientist — moderate DSA, heavy fundamentals

Data scientist loops add a coding round that's real but rarely brutal: arrays, strings, hash maps, basic recursion, and complexity, alongside ML concepts, probability, and statistics. Think "comfortable medium" rather than "LeetCode hard." The differentiator is usually your ML and stats depth and your ability to reason about data, not exotic algorithms.

• Go deep: arrays, strings, hash maps, basic recursion, Big-O reasoning, plus the probability, stats, and ML the role actually tests.
• Relevant: sorting, two pointers, sliding window — they appear, just less than in an SDE loop.
• Optional: graphs and DP — occasional; learn them after the core.

ML / Data Engineer — basically an SDE loop

This is the trap. If the title says engineer, the DSA bar jumps to near-SDE level: trees, graphs, dynamic programming, and system design are all in scope, because you'll be writing production pipelines. If you're targeting ML Engineer or Data Engineer, prep like an SDE, not like an analyst.

The bar at a glance

Role	DSA weight	What dominates instead
Data Analyst	Low	SQL, statistics, pandas, case questions
Data Scientist	Medium	ML, probability, stats + medium coding
ML / Data Engineer	High	Near-SDE DSA + system design

So how much DSA should you actually do?

Match the effort to the title. Chasing analyst roles? Get arrays, strings, and hash maps solid, then pour the rest of your time into SQL and statistics. Chasing data scientist roles? Do the comfortable-medium core and pair it with ML fundamentals. Chasing anything with "engineer" in it? Prep the full sheet like an SDE — the 90-day placement plan is built for exactly that.

This is precisely the problem SUITS solves: pick Data Analyst, Data Scientist, or ML and the curated 370 problems reorder to that role's real weighting, so an analyst isn't grinding the same hard-DP set as an ML engineer. For the cross-role view, see the DSA roadmap by role, and to lock the shared foundation first, the most important DSA topics for interviews is the right anchor before you specialise.