HomeCheatsheetsData Analyst
Pre-Interview Cheatsheet

Data Analyst — Confidence Cheatsheet

A printable, focused refresher tuned for Data Analyst. Open the sections that matter to you and walk in confident.

Tuned for Data Analyst · Business, Finance & Analytics > Data & Analytics
  • Know the analysis workflow: define question, locate data, clean data, explore, validate, analyze, visualize, communicate.
  • Refresh SQL: SELECT, WHERE, GROUP BY, JOIN, HAVING, window functions, CTEs and aggregation traps.
  • Know Python/pandas basics: DataFrame, filtering, groupby, merge, missing values, duplicates, datatypes and plotting.
  • Understand statistics basics: mean/median, variance, correlation vs causation, sampling bias, confidence intervals and A/B tests.
  • Strong analysts start from the business decision, not from the tool.
  • ETL/ELT: extracting, transforming and loading data between systems.
  • Data cleaning: fixing missing, inconsistent, duplicated or invalid records.
  • Metric definition: exact rule for calculating a KPI.
  • Outlier: unusual value; may be error or real signal.
  • Data lineage: where data came from and how it was transformed.
  • Analysis flow: business question -> metric definition -> data source -> quality check -> analysis -> insight -> recommendation.
  • SQL check: row count before/after joins, null checks, duplicate keys, aggregation level.
  • Visualization rule: choose chart based on comparison, trend, distribution or relationship.
  • Good output: answer, evidence, caveats, action.
  • How do you clean a messy dataset?
  • Explain a JOIN and common mistakes.
  • How would you investigate a drop in conversion rate?
  • What is correlation vs causation?
  • Describe a dashboard you built and the decision it supported.
  • Jumping directly to code without clarifying the business question.
  • Not checking data quality.
  • Making causal claims from correlation.
  • Using averages when medians or distributions matter.
  • Building pretty dashboards with unclear metric definitions.
  • Mentions validation and edge cases.
  • Can explain technical results in business language.
  • Knows enough statistics to avoid false conclusions.
  • Documents assumptions and reproducible steps.
I should sound like a decision analyst, not only a dashboard builder: define the question, validate the data, find the driver, explain the action.