Data Science Roadmap
Data Analyst
A data analyst answers business questions with existing data. The sequence is Excel, then SQL — which is the actual core skill — then enough statistics to avoid drawing false conclusions, then a BI tool such as Power BI or Tableau, and finally Python for anything the other tools cannot reach.
By the end: Answer a business question with SQL and present it so it changes a decision.
The Data Analyst Roadmap
Spreadsheets and Data Literacy~1 month
Start where the data already is — most organisations run on spreadsheets.
Excel for analysis
Formulas, lookups and pivot tables. Still the most widely used analysis tool there is.
Google Sheets
The collaborative equivalent, and usually where a stakeholder will send you something.
Clean data habits
One row per observation, consistent types, no merged cells. Fixing this later costs hours.
Reading a business question
Turn "why are sales down" into something a query can actually answer.
SQL — the Core Skill~2 months
The one thing that most determines whether you get hired.
Querying basics
Select, filter, sort and aggregate. Within a week you will beat most spreadsheet work.
Joins
Inner, left and the rest — and why a join can silently multiply your row count.
Window functions
Running totals, rankings and period-over-period. The step that marks an intermediate analyst.
CTEs and readable queries
Break a long query into named steps so it can be reviewed and reused.
Understanding the warehouse
Facts, dimensions and why the schema looks the way it does.
Query performance
Why a query is slow, and why in a cloud warehouse slow also means expensive.
Statistics for Analysts~1 month
Enough to avoid confidently reporting noise.
Summary statistics
Mean, median, spread — and knowing when an average is hiding the story.
Correlation versus causation
The single most consequential distinction in the job, and the easiest to get wrong in a meeting.
Sampling and significance
Whether a difference is real or the data being small.
A/B test reading
You may not design the experiment, but you will be asked what it means.
Cohorts and retention
The analysis pattern behind most product and growth questions.
Visualisation and BI~1 month
Make the answer obvious to someone who will look for six seconds.
Visualisation principles
The right chart for the comparison, honest axes, and less ink.
A BI tool
Power BI or Tableau — whichever your target employers already run.
Dashboard design
Answer one question per dashboard. A wall of tiles gets ignored within a fortnight.
Presenting findings
Recommendation first, caveats stated, evidence behind. Analysis nobody acts on is wasted.
Python and Next Steps~1 month
Raise your ceiling once SQL and BI are solid.
Python basics
For the questions a dashboard cannot express and the tasks you have repeated too often.
Pandas
Spreadsheet operations that scale past a spreadsheet, and repeat identically each month.
Automating a report
Turn a recurring manual process into a script. The clearest value you can demonstrate early.
Portfolio and interviews
Two or three analyses on real public data, plus heavy SQL practice.
Resume and applications
Lead with outcomes and the decisions your analysis changed, not the tools you touched.
Frequently Asked Questions
Is data analyst a good entry point into data careers?
It is the best one. The barrier is lower than data science, the work is genuinely useful from month one, and you spend every day with the organisation's real data — which is exactly the context a later move into data science or analytics engineering needs.
Do I need Python to be a data analyst?
Not to get hired, but it raises your ceiling quickly. SQL plus a BI tool covers most of the job. Python becomes the difference when a question needs a join no dashboard can express, a statistical test, or an automated repeat of something you have already done twenty times by hand.
Power BI or Tableau?
Whichever your target employers use — the skills transfer in days. Power BI is more common in Microsoft-centric organisations and is cheaper, Tableau is strong in enterprises and has the better visualisation defaults. Learning one properly is worth far more than sampling both.
Is Excel still relevant?
Very. It remains the most widely used analysis tool in existence, and the format your stakeholders will send you data in and expect answers back in. Pivot tables, lookups and clean modelling are still a daily skill, not a legacy one.