Career roadmap
Data Analyst
Answer the questions the business is actually asking, quickly and correctly, in a form people can act on.
Before you start Data Analyst
- Comfort with spreadsheets
- Willingness to learn SQL properly
- Clear written communication
SQL and spreadsheets
SQL is the job. Everything else is presentation.
The single most valuable skill in this role, and the first interview filter.
- SELECT, WHERE, GROUP BY, HAVING
- Inner, left and self joins
- Subqueries and CTEs
- NULL handling and its surprises
Window functions separate analysts who get promoted from those who do not.
- Window functions: rank, lag, running totals
- Cohort and retention analysis
- Funnel and conversion queries
- Date handling and time zones
Still the most widely used analytics tool in the world. Do not skip it.
- Lookup functions and pivot tables
- Power Query for repeatable cleaning
- Modelling and scenario analysis
- When to move out of a spreadsheet
Most of the job. Interviews ask how you handle dirty data specifically.
- Deduplication and identity resolution
- Handling missing and inconsistent values
- Type and format normalisation
- Documenting cleaning decisions
A wrong number in a dashboard destroys trust in every number.
- Sanity checks and reconciliation
- Row count and total validation
- Understanding source system quirks
- Escalating data problems
BuildAnswer twenty real business questions against a public database, each with the query and the answer.
Visualisation and dashboards
Getting the answer is half the job. Making it usable is the other half.
Chart choice and honesty. Interviewers show you a bad chart and ask what is wrong.
- Matching chart type to question
- Colour, scale and axis honesty
- Reducing chart clutter
- Accessibility in charts
Power BI, Tableau or Looker. Pick the one your target employers use.
- Data modelling inside the tool
- Calculated fields and measures
- Filters, parameters and drill-down
- Performance of large dashboards
Most dashboards are never opened twice. Designing for actual use is the skill.
- Designing around a decision
- Layout and information hierarchy
- Self-service versus curated views
- Maintenance and deprecation
Two teams with different definitions of 'active user' is a classic organisational failure.
- Writing unambiguous metric definitions
- Certified versus ad-hoc metrics
- Metric governance and ownership
- Handling definition changes over time
The output is a decision. Structure the message accordingly.
- Leading with the answer
- Writing for executives
- Presenting uncertainty honestly
- Handling challenges to your numbers
BuildA dashboard a real person uses weekly, with a documented definition for every metric.
Analysis techniques
Move from reporting what happened to explaining why and what to do.
Enough to avoid confidently reporting noise as a finding.
- Averages, medians and distribution shape
- Variability and confidence intervals
- Correlation and its limits
- Significance in plain language
Aggregates hide the story. Segmentation is where insight usually lives.
- Cohort analysis over time
- Behavioural segmentation
- RFM and customer value analysis
- Simpson's paradox in segments
The most common analysis type at product companies.
- Funnel definition and drop-off analysis
- Retention curves
- Event tracking and instrumentation gaps
- Attribution basics
Business planning needs numbers for next quarter, not last one.
- Trend and seasonality
- Simple forecasting methods
- Scenario and sensitivity analysis
- Communicating forecast uncertainty
Analysts are often asked to interpret tests they did not design.
- What a p-value does and does not say
- Sample size adequacy
- Spotting a broken experiment
- Practical versus statistical significance
BuildA deep-dive analysis explaining a real trend, with alternative explanations considered and ruled out.
Scaling your work
Stop being a query service. Build things that answer questions without you.
The step up from spreadsheets, and the gateway to further roles.
- pandas for analysis
- Automating repetitive reports
- APIs and data collection
- Notebooks to scheduled scripts
Understanding the warehouse makes your queries correct and fast.
- Fact and dimension tables
- Grain and joins that do not fan out
- Slowly changing dimensions
- Reading a data dictionary
The bridge into analytics engineering, and increasingly expected in analyst job ads.
- Models, refs and lineage
- Tests for data quality
- Documentation generation
- Version control for analytics
Every manual weekly report is an hour you never get back.
- Scheduled refreshes and alerts
- Parameterised reports
- Email and chat delivery
- Monitoring for silent failures
Turning a vague request into an answerable question is most of the value you add.
- Clarifying the real question
- Managing request queues and priorities
- Saying no to low-value work
- Following up on impact
BuildAutomate a recurring report end to end and document the model behind it.
Interview preparation
Analyst interviews are SQL, a case, and a presentation of past work.
Live SQL against a schema you have just been shown. Practice under time pressure.
- Joins and aggregation quickly
- Window functions
- Debugging a query that returns wrong rows
- Explaining your query as you write
Metrics dropped last week. What do you look at, in what order.
- Structured diagnosis of a metric change
- Segmenting to isolate the cause
- Ruling out data quality issues
- Recommending next steps
Present an analysis to a mixed audience and defend it.
- Structuring a ten-minute analysis talk
- Anticipating challenges
- Admitting limitations well
- Handling a hostile question
Expect direct questions on whichever BI tool the job uses.
- Power BI DAX or Tableau calculations
- Data model design in the tool
- Performance troubleshooting
- Row-level security
Public dashboards and write-ups substitute for experience.
- A public dashboard with real data
- Written analyses with recommendations
- Path to analytics engineering
- Path to data science
BuildA portfolio of three analyses with the question, method, answer and recommendation.
Data Analyst tools on your CV
- SQL
- Excel / Sheets
- Power BI / Tableau
- Python
- dbt
- Looker
- BigQuery / Snowflake
What Data Analyst employers ask to see
- A public dashboard someone actually uses
- Three written analyses with recommendations
- An automated recurring report
- A documented metric definition set
The most accessible entry point into data work, hiring across every industry rather than only tech. Progresses naturally into analytics engineering or data science.
Content last reviewed 2026-08-31. Guidance only — no institute or paid placement is endorsed anywhere in this book.