Practical learning and project workspace

Free Roadmaps

Free Data Analyst Roadmap: Foundations, Tools, and a Practical Portfolio

A two-phase data analyst learning guide covering Excel, SQL, Python, Power BI, Tableau, a flexible twelve-week plan, portfolio work, and Pakistan and global applications.

DataUpto Editorial13 min read2,556 wordsResearch checked 27 October 2026

Free to learn. Build at your own pace.

The guide and linked videos are free to access. Software plans, certificates, advertising, API calls and cloud hosting may cost money. Use the official resource notes before signing up. Courses are curated for topic fit, not a verified ratings leaderboard.

1. What this free data analyst roadmap is designed to achieve

A data analyst turns imperfect records into useful decisions. That means defining a question, checking whether the available data can answer it, calculating defensible metrics, and explaining the result to someone who must act. Software matters, but a beautiful dashboard with the wrong revenue definition is still wrong. This roadmap therefore pairs every tool with a business question, an exercise, and a completion checkpoint.

There are two phases: foundations, followed by intermediate and advanced application. The twelve-week schedule is an example for someone able to study consistently, not a job guarantee or a universal deadline. If you are new to computers, English documentation, or basic mathematics, repeat weeks rather than rushing. A useful study session includes reading, building without copying, and writing down what failed. Your goal is independent reasoning, not collecting course completion badges.

2. Free learning, software choices, and paid-platform caveats

Start with official documentation and public datasets whose licenses permit your intended use. Python, pandas, PostgreSQL, and SQLite can support local practice without subscription fees. Excel licensing depends on the edition and access available to you; its browser version and desktop versions do not have identical features. If you lack Excel access, practice spreadsheet reasoning in LibreOffice Calc or a browser spreadsheet, then verify Excel-specific menus and formulas when access becomes available.

Power BI Desktop is free for authoring, while service-based sharing and collaboration depend on licenses and organizational capacity. Check Microsoft's current licensing documentation before promising a client a free shared dashboard. Desktop also requires a supported Windows environment. Tableau Public is appropriate only for material safe to publish publicly; it is not a confidential-data hosting solution. Paid learning sites may charge for assessments, certificates, labs, or continued access even when lessons are advertised as free. Check terms before entering payment details.

3. Phase one: build a dependable spreadsheet foundation

Learn rows, columns, cell references, data types, rectangular tables, and the difference between a blank cell and a genuine zero. Practice relative and absolute references, sorting, filters, freeze panes, and data validation. Then learn SUM, AVERAGE, COUNT, COUNTA, IF, SUMIFS, COUNTIFS, IFERROR, and date functions. Use XLOOKUP where supported and understand an alternative lookup approach for older workbooks. Avoid memorizing formulas without knowing what each input represents.

Create a fictional order table containing order date, customer identifier, product, units, price, discount, and region. Calculate net revenue using an explicitly documented discount convention. Introduce inconsistent capitalization, missing regions, and dates stored as text, then clean them without destroying the original sheet. Build a pivot table showing monthly revenue by region. Your checkpoint is to reconcile the pivot total with a separate formula calculation and explain why averaging row-level percentages may give a misleading overall margin.

A collaborative learning workspace — Editorial illustration · Build skills through practice, not video completion alone.

4. SQL foundations: ask precise questions of relational data

Use SQLite for a simple local start or PostgreSQL for a fuller database environment. Learn SELECT, aliases, WHERE, ORDER BY, LIMIT, DISTINCT, CASE, aggregation, GROUP BY, and HAVING. Understand primary keys, foreign keys, and table grain: one row might mean an order, an order line, or a customer. Learn INNER JOIN and LEFT JOIN before experimenting with more complicated combinations. Treat NULL as missing or unknown information rather than assuming it behaves like zero.

Split your spreadsheet into customers, orders, and order_items tables. Write queries for monthly revenue, customers with no orders, and the five largest product categories. Before every join, predict the expected row count and identify whether the relationship is one-to-one or one-to-many. Your checkpoint is to detect an intentionally duplicated customer key, explain how it inflates totals, and correct the model rather than hiding the problem with DISTINCT. Save both the SQL and a short description of each query's business purpose.

5. Python foundations and reproducible analysis

Learn variables, numbers, strings, lists, dictionaries, conditions, loops, functions, imports, and basic exception handling. Practice reading a file path, checking whether a file exists, and understanding a traceback. Use a virtual environment and record dependencies so another person can recreate your setup. A notebook is useful for exploration, but it should run from top to bottom after restarting its kernel; hidden execution state makes results difficult to trust.

Write a small script that reads a CSV, checks required column names, converts units and prices to numbers, and reports invalid rows. Separate loading, cleaning, and calculation into functions rather than one long block. Learn enough pandas to inspect shape, types, missingness, and duplicates, then filter rows, create columns, group records, and merge tables. Your checkpoint is to reproduce one spreadsheet total in Python, investigate any difference, and export a clean summary without manually editing the output file.

6. Data quality, statistics, and ethical interpretation

Learn mean, median, percentiles, range, variance, standard deviation, proportions, and weighted averages. Understand why a skewed distribution may need a median as well as a mean. Distinguish a population from a sample, correlation from causation, and a descriptive pattern from a reliable prediction. Learn the intuition behind confidence intervals and sampling variability without treating a single threshold as proof that a business change worked.

Create a quality report covering missing values, duplicate keys, impossible dates, negative quantities, and suspicious outliers. Decide which rules reject a record and which merely flag it for review. Do not automatically delete unusual customers or large transactions: they may be valid and important. Your checkpoint is a one-page analysis comparing two regions while acknowledging different customer mixes and sample sizes. Remove personal identifiers from public work, use licensed data, and document how anonymization can still leave re-identification risks when detailed attributes are combined.

7. Phase two: intermediate SQL and analytical Python

Move into common table expressions, subqueries, window functions, conditional aggregation, and date-based analysis. Practice ROW_NUMBER, RANK, LAG, and running totals, paying attention to partitioning and ordering. Learn how a date spine helps represent months with no sales instead of silently dropping them. Understand that SQL dialects differ in date functions and syntax, so check the documentation for your chosen database rather than pasting queries blindly.

In pandas, practice merge validation, pivot tables, datetime operations, string cleaning, and explicit handling of missing categories. Create a customer cohort table and calculate repeat-purchase behavior using a clear observation window. Compare a SQL result with a pandas result as a cross-check. Your checkpoint is to explain whether an apparent decline comes from fewer customers, fewer orders per customer, or lower order value. Add small tests for known examples and document how refunds and canceled orders affect each metric.

8. Power BI: build a semantic model before decorating charts

Learn importing CSV and database data, Power Query transformations, data types, and refresh behavior. Build a star schema with fact and dimension tables, a proper date table, and relationships you can explain. Distinguish calculated columns from measures. Start DAX with SUM, COUNTROWS, DISTINCTCOUNT, DIVIDE, and CALCULATE, then study filter context before attempting complex time intelligence. Know which transformations belong upstream instead of repeating them inside every visual.

Build a sales report with a summary page, a regional comparison, and a product detail page. Add date and region slicers, readable labels, and a clear definition for revenue. Test totals with no filters, one region selected, and multiple months selected. Your checkpoint is to reconcile every headline measure against SQL and explain why a measure changes when a slicer changes. Use screenshots or an appropriate public demonstration when paid sharing is unavailable, and never use public publishing as a workaround for protecting confidential information.

9. Tableau: a second visualization language, not a duplicate obsession

Choose one BI tool as your primary portfolio tool and learn the other well enough to understand its workflow. In Tableau, practice dimensions and measures, discrete and continuous fields, aggregation, calculated fields, filters, sorting, and dashboard actions. Learn how relationships and joins affect analysis. Start with bars, lines, and scatterplots; maps are useful only when location is genuinely relevant to the question.

Recreate one Power BI page in Tableau Public using synthetic or properly licensed public data. Add a trend chart, a category comparison, and a detail view connected by an understandable interaction. Check labels, tooltips, color contrast, and behavior on a smaller screen. Your checkpoint is to explain one design choice you changed between tools and why it improves comprehension. Assume published Tableau Public content is public, review its visibility and download settings, and exclude names, private client records, credentials, and sensitive business assumptions from the workbook.

A practical project workspace — Editorial illustration · Build skills through practice, not video completion alone.

10. Communication: turn a finding into an actionable recommendation

Before analyzing, write a question such as, “Which product groups contributed to the quarterly decline, and what should operations investigate first?” Define the audience, decision, metric, time range, and exclusions. A useful recommendation connects an observed pattern to an action and a way to check whether the action helped. Avoid claiming that a descriptive dashboard has proved the cause of a change.

Practice a five-minute presentation with four parts: question, evidence, limitation, and next step. Replace “sales are bad” with a precise statement about the metric and period, while separating an invented exercise scenario from a real business claim. Write a short executive summary and a technical appendix so different readers can inspect different levels of detail. Your checkpoint is to give your dashboard to a friend without guidance, ask what decision it supports, and revise any chart whose meaning they cannot explain. State uncertainty directly rather than hiding it behind jargon.

11. A realistic twelve-week example plan

Weeks one and two cover spreadsheet tables, formulas, cleaning, and pivot tables; finish with a reconciled regional sales summary. Weeks three and four cover SQL selection, aggregation, joins, and missing values; finish with a small database and a documented query set. Weeks five and six introduce Python, pandas, descriptive statistics, and quality checks; finish by reproducing the spreadsheet analysis in code.

Weeks seven and eight focus on intermediate SQL, customer cohorts, and a Power BI data model. Weeks nine and ten add a usable dashboard, a smaller Tableau exercise, and a written stakeholder explanation. Weeks eleven and twelve are for portfolio integration, peer feedback, interview practice, and targeted applications. A possible weekly rhythm is three study sessions, two independent exercises, and one review session. If you cannot complete the checkpoint unaided, repeat the week. Someone with limited study time may sensibly stretch this example over several additional months.

12. Portfolio project: scope an end-to-end retail analysis

Build a fictional retail performance review or choose an openly licensed retail dataset. Define three questions: what changed over time, which segments contributed, and which operational follow-up is justified. Write a data dictionary specifying the grain of each table, the currency, date conventions, refund treatment, and whether revenue includes tax or shipping. If the dataset lacks costs, do not pretend to calculate profit; analyze revenue and clearly label that limitation.

Step one is to preserve the original files unchanged and record their source and license. Step two is to profile columns, missing values, ranges, and keys. Step three is to clean data using repeatable SQL or Python transformations, keeping a log of rejected records. Step four is to create fact and dimension tables and validate relationship cardinality. Your checkpoint is a reconciliation sheet showing raw totals, excluded amounts, and final totals, with a reason for each difference. This evidence is more useful than an unexplained polished screenshot.

13. Portfolio project: analysis, dashboard, and handover

Step five is to write SQL for monthly performance, product contribution, repeat purchasing, and regional comparisons. Step six is to reproduce selected calculations in pandas and create diagnostic charts. Step seven is to build a Power BI or Tableau dashboard with a visible date range, metric definitions, sensible default filters, and no unnecessary decoration. Include an explicit “data last updated” note rather than implying a static file is live.

Step eight is to write a one-page decision memo containing the finding, supporting evidence, plausible alternative explanations, and a next investigation. Step nine is to publish a clean repository with a README, setup instructions, SQL, scripts, screenshots, and sample outputs. Keep secrets and restricted raw data out of version control. Step ten is to ask another person to reproduce a result and record what you fixed after their feedback. Completion means someone can understand your question, rebuild the permitted analysis, and see where your confidence ends without contacting you for missing instructions.

14. Applying in Pakistan and competing for global work

In Pakistan, search beyond the exact title “data analyst”: reporting analyst, MIS analyst, operations analyst, business intelligence intern, and finance reporting roles may use overlapping skills. Read individual requirements rather than assuming every employer uses the same stack. Check company career pages and established job boards, and use university groups or local professional communities for introductions. Tailor your resume to the work you can demonstrate, with concise project descriptions rather than unsupported claims of expertise.

For international remote opportunities, check country eligibility, time-zone expectations, communication requirements, and whether the arrangement is employment or contracting. Do not assume a remote listing accepts applicants everywhere. For freelance work, define deliverables, data access, revision limits, acceptance criteria, and payment terms before starting. Demonstrate clear written English through a short portfolio memo, while using Urdu or other local languages when they help stakeholders understand the work. Never pay a recruiter to unlock a supposed guaranteed position, and verify organizations before sending identity documents or production data.

15. Interview preparation and common questions

What should you practice for SQL interviews? Explain joins, aggregation, NULL handling, and window functions while narrating the grain of the result. Solve a small question first, then test edge cases such as duplicated keys and customers without orders. For spreadsheets, practice lookup errors, pivot tables, and a weighted average. For BI interviews, explain relationships, measure definitions, filter behavior, and how you checked a dashboard's accuracy.

Do you need both Power BI and Tableau immediately? No: depth in one plus basic familiarity with the other is usually a better learning strategy than two shallow portfolios. Do you need Python before applying? Requirements vary; apply where your demonstrated strengths match the role, while continuing to learn. Is a certificate enough? It can document study, but it does not replace reproducible work. Can this plan guarantee a job in three months? No. Hiring depends on role fit, prior experience, communication, location, eligibility, and employer demand, none of which a study schedule controls.

16. Completion checklist and what to learn next

You have a useful foundation when you can take an unfamiliar table, identify its grain, diagnose quality problems, join it safely, calculate a defensible metric, and explain the result without overstating the evidence. Review your portfolio against that standard rather than counting videos watched. Keep a learning log with the mistake, the cause, the correction, and the check that would catch it next time. This habit makes interview explanations more credible and future work more reliable.

After finishing, choose a direction based on actual job descriptions: deeper Power BI modeling, product analytics and experimentation, finance reporting, marketing analysis, or analytics engineering. Learn Git basics and automated validation before adding more dashboard tools. Revisit your project periodically to improve documentation and remove fragile manual steps. Use the official resources below as references, not as a requirement to read every page in order. The strongest next exercise is a slightly unfamiliar question you can answer independently while showing both your calculations and their limitations.

freeCodeCamp.org

Version control, branches and portfolio repositories; GitHub account needed to publish.

Watch on YouTube

LESSON 11

SQL / Excel / Tableau / Power BI / Python

Data Analyst Bootcamp for Beginners (SQL, Tableau, Power BI, Python, Excel, Pandas, Projects, more)

freeCodeCamp.org

Integrated portfolio projects; overlaps earlier foundations, so use as capstone rather than repeating everything.

Watch on YouTube

Free resources & official downloads

Excel free web app

Free browser edition requires account; desktop apps and some advanced capabilities are paid.

Power BI Desktop download and installation

Free Desktop authoring on Windows; not a native macOS application.

Power BI licensing

Free authoring does not imply free private organizational sharing; sharing depends on licenses/capacity.

Tableau Public download

Free; registration may be required. Published work is public: never upload confidential data.

Tableau sample datasets

Free practice data; check individual dataset terms.

Python downloads

Free open-source interpreter; choose a supported release compatible with libraries.

PostgreSQL downloads

Free local database; MySQL-based course syntax may need adaptation.

pandas documentation

Free getting-started guides and API reference; older video APIs may differ.

Explore free datasets

Sources & further reading

Career examples are guidance, not vacancy announcements or job guarantees. Confirm requirements, regional access and current product terms at the linked official source.

Choose your next practical step

Ask Qasim Anwar for guidance on your learning path, projects or career applications.

A detailed two-phase digital marketing roadmap covering research, copy, SEO, analytics, Google and Meta campaign planning, consent-aware email, CRM, experimentation, and an honest no-spend portfolio.

15 min read10 video resourcesRead guide