Recruiters ask for a portfolio. Most beginners do not have one. This is how to build one that actually works from nothing, in under four weeks.

Here is the problem most beginners run into. They finish the course. They know SQL and Python well. Then, they apply for analyst roles but hear nothing back. Many times, it’s one of two reasons: their resume lacks proof of real work, or their GitHub is empty when a recruiter checks.

A portfolio solves both problems. It is concrete, shareable evidence that you can do the work. Not that you passed a quiz at the end of a course. Not that you watched forty hours of tutorials. You used a real dataset and asked a genuine question. Then, you cleaned the data and built an analysis. Finally, you shared the findings clearly.

This guide shows you how to build your portfolio. It covers what to include, where to host it, how to write about your projects, and how to present it. This way, a recruiter can understand it in just thirty seconds.

Step 1: Understand What a Portfolio Actually Is (And What It Is Not)

A data analyst portfolio is not a collection of course completion screenshots. It is not a Kaggle notebook you followed step by step from a tutorial video. It is not a GitHub repository with three files and no explanation of what any of them do.

A portfolio includes three to five projects. Each project highlights a specific skill. This could be data cleaning, SQL querying, visualization, statistical analysis, or end-to-end reporting. Each project must be documented clearly. A recruiter who hasn’t seen your work should easily open it. They need to grasp the problem you solved, your approach, and your findings.

Think about this before adding a project: If a recruiter clicks the link, will they understand your work and skills in two minutes? If the answer is no, the project needs more documentation, not more code.

• Include: Projects you built yourself, on real or realistic datasets, with clear documentation
• Include: Projects where you made actual decisions what to clean, what to analyse, how to visualise
• Skip: Copied tutorials where every step was given to you in a video
• Skip: Course exercises and challenge solutions with no original thinking

Step 2: Choose Three to Five Projects That Cover Different Skills

Three strong projects beat ten weak ones every time. The goal is not volume. The goal is coverage showing that you can handle different types of data work, not just one thing repeatedly.

A well-rounded beginner portfolio covers at least three of these five skill areas:

• SQL: A project where the core work is writing queries – joins, aggregations, window functions, CTEs
• Python + EDA: A project where you clean a messy dataset, explore it, and produce visualisations that explain what you found
• Dashboard / BI tool: A published Power BI report or Tableau Public dashboard built from a real dataset
• Statistics: A project that involves hypothesis testing, correlation analysis, or regression even a simple one
• End-to-end storytelling: A project that goes from raw data to a clear business recommendation, written up as if presenting to a non-technical stakeholder

You do not need all five. Three that are documented well will get more attention than five that are not. If you covered the Python EDA project and the SQL project from the beginner projects list, you have two. Add one BI dashboard and you have a complete, defensible portfolio.

Step 3: Host Everything on GitHub

GitHub is where recruiters and hiring managers look when they want to see your work. A portfolio that lives only in a local folder on your laptop does not exist as far as the job market is concerned. Every project needs a GitHub repository.

Setting up a GitHub account takes ten minutes. Creating a repository for a project takes two. If you have never used Git before, learn the five commands you need — git init, git add, git commit, git push, git clone — and nothing else for now. Git mastery is not the goal. Hosting your work publicly is.

Each repository needs three things to be useful:

1. A README file that explains the project in plain English
2. The actual code or queries, clean and commented
3. Either the dataset itself or a clear link to where it can be downloaded

The README is the most important file in the repository. It is what a recruiter reads. Most beginners skip it or write two lines. Write a proper one.

Step 4: Write a README That Does the Selling for You

A good project README has five sections. It does not need to be long. It needs to be clear.

• Problem statement: What question were you trying to answer? One to two sentences, written as if explaining to someone who does not know what your dataset is.
• Data source: Where the data came from, how many rows, what the key variables were. Include a link if it is publicly available.
• Tools used: Python, pandas, matplotlib, SQL, Power BI — list what you actually used in this project
• Approach: What did you do? Three to five bullet points covering the main steps — loaded the data, cleaned nulls and inconsistent categories, built aggregations, visualised distributions, identified the key finding
• Key findings: Two to four bullet points on what you actually discovered. Write these as conclusions, not as descriptions of charts. Not ‘a bar chart showing sales by region’ but ‘the North region underperformed by 23% relative to target in Q3, driven by a 40% drop in electronics sales’

Weak README — what most beginners write This project analyses sales data using Python. I used pandas and matplotlib. The dataset has information about sales and customers.

 

Strong README — what gets recruiters to read further This project analyzes over 100,000 e-commerce transactions from the Olist dataset. It identifies which product categories underperformed in Q3. It also examines if delivery delays affected customer review scores. Tools used include Python, pandas, matplotlib, and seaborn. Delivery delays over 10 days cut average review scores by 28% in all categories. Electronics saw the biggest drop in sales, but its percentage decline was third. This makes the headline metric misleading.

 

Step 5: Add a Tableau Public Dashboard or Power BI Published Report

GitHub hosts your code. Tableau Public and Power BI’s sharing features host your dashboards. Both are free. Both are publicly accessible via a link. A recruiter who clicks a Tableau Public link sees your actual dashboard running in their browser — no download, no setup, no risk of formatting breaking.

Build one dashboard on Tableau Public or one report published to the Power BI Service. Use the same dataset you used for one of your Python projects if possible — showing the same data from two different angles (code-based analysis and a dashboard) demonstrates more range than two completely separate projects.

When publishing to Tableau Public, write a description for the workbook that includes: what business question it answers, what data was used, and what the key takeaway is. Recruiters browsing Tableau Public will find your work through search — a descriptive title and summary increases that chance.

• Tableau Public: Free hosting, publicly accessible dashboard link, searchable by topic
• Power BI Service: Free for 60 days with a work or school email, shareable via link
• Tip: Publish the dashboard, copy the link, and add it directly to the GitHub README of the same project

Step 6: Create a Portfolio Summary Page

Once you have three projects on GitHub and at least one dashboard live on Tableau Public or Power BI, you need one page that ties everything together. This is the single link you put on your resume and LinkedIn profile.

The options are:

• GitHub Profile README: GitHub lets you create a special repository with your username as the repo name. A README in this repository appears on your GitHub profile page. Use it as your portfolio index — a brief introduction, a list of your projects with one-line descriptions and direct links, and the skills each project demonstrates.
• Notion or a personal website: A simple Notion page works well as a portfolio page. Add a project card for each piece of work with a thumbnail, a two-line description, the tools used, and a link to the GitHub repo and live dashboard. Free to set up and looks significantly more polished than a plain GitHub profile.
• LinkedIn featured section: Pin your best project to the Featured section of your LinkedIn profile. Write a one-paragraph post about it with a link to the GitHub repo. Projects shared as LinkedIn posts get significantly more visibility than links buried in a bio.

The goal is one URL that represents all your work. When you write ‘portfolio: github.com/yourname’ on your resume, a recruiter who clicks it should immediately understand what you have built and at what level.

Step 7: Link It Correctly on Your Resume and LinkedIn

A portfolio that is not visible does not help you. The link belongs in three specific places:

• Resume header: Alongside your phone number and email — ‘Portfolio: github.com/yourname’ or your Notion link. Not buried at the bottom of the page.
• LinkedIn summary: The first line of your LinkedIn About section should mention your portfolio and include the link. This is visible without expanding the section.
• Project bullets on your resume: Every project bullet point that describes a piece of portfolio work should end with the direct link to that specific repository or dashboard. Not the general GitHub profile — the specific project.

Wrong — generic link, no context Built data analysis projects. GitHub: github.com/yourname

 

Right — specific link, specific project E-Commerce Sales Analysis | Python, pandas, Power BI — Analysed 100K+ transactions, found delivery delay impact on ratings. github.com/yourname/ecommerce-sales-analysis

 

The One Mistake That Kills Most Portfolios

Building projects but not writing about them.

Seventy percent of beginner portfolios have decent code and almost no explanation. The README is two lines. The Tableau dashboard has no description. The GitHub profile has no summary. A recruiter who lands on these repositories has no idea what they are looking at, why it matters, or whether the person who built it understood what they found.

Writing is not separate from the portfolio. It is half of it. An analyst who cannot explain their work clearly in writing cannot explain it in a stakeholder presentation either. Recruiters know this. A well-written README with clear findings is itself a demonstration of the communication skill every analyst job description asks for.

Spend as much time on your documentation as you spend on your code. It is not overhead. It is the part that gets read first.

Wrapping Up

A portfolio that lands interviews isn’t big. It’s specific, well-organized, and easy to explore. Here’s the full checklist:

• Three to five projects in various skill areas:

  • SQL

  • Python EDA

  • BI dashboard

  • End-to-end analysis

• Every project hosted on a public GitHub repository with a proper README

• At least one dashboard live on Tableau Public or published via Power BI

• A portfolio summary page GitHub profile README, Notion, or personal site

• Include a portfolio link in the resume header.

  • Add it to the LinkedIn summary.

  • Place it next to each relevant project bullet.

• Each README has:

  • Problem statement

  • Data source

  • Tools

  • Approach

  • Key findings as conclusions

Four weeks of focused work on three projects can improve your chances for entry-level analyst roles. Make sure to document your work well and share it publicly. Candidates who get interviews aren’t always the most skilled. They are often the ones whose work is easiest to assess.

Read Also:

A Day in the Life of a Data Analyst in India

Excel vs Power BI vs Tableau for Data Analysts (2026 Guide)

20 Highest Paying Companies Hiring Freshers in India (2026)

Job Notification Join us on Telegram: Click here

Job Notification Join us on WhatsApp: Click here