Data science combines statistics, domain knowledge, and computation to extract insight from data—often iteratively, not in a single linear pipeline.
Core activities
- Ask — define measurable questions and success metrics
- Acquire — databases, APIs, files, experiments
- Clean — types, missing values, duplicates, leakage
- Explore — distributions, relationships, anomalies
- Model — predict, classify, cluster (when appropriate)
- Communicate — charts, narratives, recommendations
Roles you will see
- Analyst — SQL + dashboards + stakeholder updates
- Data scientist — experimentation, modeling, evaluation
- ML engineer — production pipelines and serving
Skills overlap—this track builds the shared foundation.
Important interview questions and answers
- Q: CRISP-DM idea?
A: Common process: business understanding → data → modeling → evaluation → deployment. - Q: Always need deep learning?
A: No—many wins come from clean data, good features, and simple models.
Self-check
- Name the six activities in order.
- One role that emphasizes SQL and dashboards?
Tip: Sketch the six-step loop on paper for your current project.
Interview prep
- Six steps?
Ask, acquire, clean, explore, model, communicate.
- Always ML?
No—many wins are clean data and clear metrics without models.