Sefism Data Analytics Roadmap

What this roadmap is

The Data analytics roadmap is Sefism's path for students and early-career engineers who want to turn data into decisions as a data or BI analyst. It starts where good analysts start, with a sharp business question and precise metric definitions, then builds the skills every analyst uses (spreadsheets, SQL and statistics) before going deep on one analysis tool, cleaning and modelling real data, and presenting a clear recommendation.

When you set it up, you choose your tool:

• Excel and Power BI: the most common in business teams.
• Tableau: strong visual analytics, with a free Public edition.
• Python: pandas and notebooks, the most flexible.

Spreadsheets and SQL are taught to everyone, whatever tool you choose, because they are what the rest of a business checks your numbers against. Stage 5 becomes Power BI in depth, Tableau in depth or Python for analysis, and the projects shift slightly with your choice.

What you can do at the end

• Turn a vague request into a specific business question with written metric definitions.
• Clean and summarise data in Excel or Google Sheets with lookups, conditional aggregation and pivot tables.
• Write SQL with joins, CTEs and window functions for rankings, running totals and cohorts.
• Explain averages, spread, correlation and an A/B test result with a confidence interval in plain words.
• Build a trustworthy model and dashboard in Power BI, Tableau or Python that reconciles with an independent total.
• Present a finding as a one-sentence message with clean charts, and publish two or three portfolio projects.

Who it suits, and who should start elsewhere

It suits students from any degree, early-career engineers, and people in finance, operations or marketing who already work with spreadsheets and want to move into an analyst role. You do not need a statistics degree; stage 4 covers what analysts use every week.

If you want to build models and predictions rather than answer business questions, look at Data science or Machine learning. If you enjoy the SQL and pipeline side most and want to build the data infrastructure analysts rely on, Data engineering is the next step. For deeper database design and performance than stage 3 covers, follow SQL and databases.

Time and pace

The eight stages total about 238 hours, the same for every tool. The first stage, How analysts think, is a short foundation you can skim if you already work with business questions.

A student's week (10 to 12 hours): write SQL by hand on most days (the roadmap suggests every day for a month in stage 3), one longer session in your tool on a public dataset, and one session writing up what you found. Reuse the same dataset across stages so your numbers can be checked against each other.

Someone working full time (about 5 hours): three short SQL sessions in the week and one weekend block on your tool. If your job already involves spreadsheets, stage 2 will go quickly; use the saved time on SQL.

Stage by stage

1. How analysts think (about 8 hours, foundation)

What a data analyst does all day, asking the right business question, metrics, KPIs and precise definitions, descriptive, diagnostic, predictive and prescriptive analysis, the analysis workflow and tidy data, and personal data, privacy and responsibility.

Milestone: a one-page analysis brief for a question you care about: the question, who would use the answer, three metric definitions with edge cases, the type of analysis and where public data could come from.

2. Spreadsheets properly (about 25 hours)

References and structured tables, SUMIFS, COUNTIFS and IF logic, XLOOKUP, cleaning text, numbers and dates, pivot tables for exploring, and a workbook other people can trust.

Milestone: a documented workbook from a public dataset of at least 5,000 rows, with raw, working and summary tabs, a lookup from a second table, a SUMIFS grid, two pivot tables and a notes tab listing every cleaning step.

3. SQL for analysis (about 35 hours)

SELECT, WHERE and how a query runs, GROUP BY and HAVING, joins without inflating totals, NULLs and CASE WHEN, CTEs and subqueries, window functions for ranking and top N, and running totals, LAG and period comparisons. The roadmap notes that SQL is the skill analyst job descriptions ask for most.

Milestone: at least ten saved queries on a multi-table public dataset, each answering a named business question, including joins with row-count checks, a cohort query with CTEs, a top N per group and a month-on-month change with LAG.

4. Statistics you need (about 25 hours)

Averages and spread, distributions and outliers, samples and sampling bias, correlation, causation and Simpson's paradox, A/B test basics, and confidence intervals and p-values at the level of intuition.

Milestone: a statistics notebook or workbook with summaries and percentiles, two commented histograms, a correlation with a note on confounders, and a public A/B test analysed with a confidence interval and a plain-language conclusion.

5. Your analysis tool in depth (about 50 hours)

• Power BI: Power Query for repeatable preparation, the star schema, DAX measures, CALCULATE, filter context and time intelligence, building reports, and publishing and sharing. Milestone: a refreshable report with a marked date table and at least eight DAX measures, committed to GitHub with a README.
• Tableau: relationships and joins, dimensions and measures, calculated fields, LOD expressions and table calculations, dashboards, actions and parameters, and Tableau Public. Milestone: a Tableau Public workbook with at least six calculated fields, including a FIXED LOD and a table calculation, and an interactive dashboard.
• Python: Jupyter and your setup, loading, inspecting and selecting in pandas, cleaning, groupby and merge, charts with matplotlib and seaborn, and notebooks as reports. Milestone: a report notebook with counted cleaning changes, validated merges, groupby results that match your SQL, and four polished charts.

6. Cleaning and modelling data (about 25 hours)

Profiling a new dataset, missing values (decide, do not default), duplicates, keys and consistent categories, types, dates and parsing, grain, facts and dimensions, ratios, rates and correct totals, and reconciling totals with a cleaning log.

Milestone: a genuinely messy public dataset cleaned and modelled in your tool, with a profiling summary, a cleaning log, stated grain for every table, and two headline totals reconciled against an independent figure.

7. Charts, dashboards and storytelling (about 30 hours)

Choosing the right chart, clutter, colour and focus, accessible and honest charts, designing a dashboard people use, the one-sentence message, structuring the story, and presenting to stakeholders and handling questions.

Milestone: a finished dashboard or report with takeaway titles and accessible design, plus a five-slide presentation or one-page memo around a one-sentence finding, presented to at least one person who gave you feedback.

8. Portfolio and first analyst role (about 40 hours)

Finding public data worth analysing, two or three question-driven projects, writing up each project, a resume and profile for analyst roles, how analyst hiring works, adjacent roles and internal moves, freelance data work done safely, and certificates, honestly.

Milestone: a public portfolio of two or three projects with one-page write-ups, linked from a one-page resume with results-first bullets, plus ten tracked applications or conversations, or one scoped freelance proposal.

Projects to build

• Online store sales and customer retention dashboard: the report almost every analyst builds at work, in any of the three tools.
• A/B test analysis report on a public experiment dataset: the statistics stage in practice.
• Public budget transparency explorer: real, messy public data to clean and model.
• Sports performance analytics from match event data: a larger pandas project for the Python route.

How it works on Sefism

The public page at /roadmaps/data-analytics/ shows the full stage list, with the first two stages open to read. Everything on Sefism is free to learn from. A few actions use credits, such as adding a roadmap to your dashboard; new members get starting credits, the first roadmap is free, and if you are learning seriously and run out you can ask for more.

Adding the roadmap opens every stage in your dashboard. You choose your tool, set weekly hours and familiarity, and tick off topics as you go; each topic has an explanation, a check question and a prompt that makes an AI assistant quiz you rather than doing the work. Following the roadmap also unlocks its companion cheat sheet. For help choosing a tool or a portfolio project, send Tauseef a question or book a 1:1 session.

The companion cheat sheet

The Data analytics cheat sheet follows the same order, with Power BI, Tableau and Python editions and formulas, SQL and pandas snippets. Use it after each stage and before an analyst technical round or take-home task.

Related on Sefism

• What is the difference between a data analyst, a data scientist and a machine learning engineer?
• How do I switch from data analyst to data engineer?
• How do I become a business analyst in tech?
• Data story from your previous field using public data, a good first project if you are switching careers.
• Tableau Public, DAX Guide and Python for Data Analysis, used in stage 5.

What you can do afterwards

You will have a workbook, a saved set of SQL queries, a statistics notebook, a modelled and reconciled dataset in your tool, a dashboard with a written recommendation, and two or three portfolio projects. That is the evidence analyst hiring rounds look for, and a base for moving later into data science or data engineering.

Frequently asked questions

How long does it take to become a data analyst? This roadmap is about 238 hours: roughly 20 weeks at 12 hours a week, or about 48 weeks at 5 hours a week.

Should I learn Power BI, Tableau or Python? Power BI is the most common in business teams (and Power BI Desktop runs only on Windows), Tableau is strong for visual analytics with a free Public edition, and Python is the most flexible. Pick one and go deep; the roadmap advises against touching all three.

Is SQL necessary for data analysts? Yes. Stage 3 calls it the skill analyst job descriptions ask for most and the one tested most often in hiring.

Do I need a degree in statistics? No. Stage 4 covers averages, spread, distributions, sampling bias, correlation and A/B tests at the level analysts use every week, with a focus on not overclaiming.