{"id":383945,"date":"2026-09-28T19:50:15","date_gmt":"2026-09-28T19:50:15","guid":{"rendered":"https:\/\/colorlib.com\/wp\/data-science-portfolio-examples\/"},"modified":"2026-09-28T19:50:15","modified_gmt":"2026-09-28T19:50:15","slug":"data-science-portfolio-examples","status":"publish","type":"post","link":"https:\/\/colorlib.com\/wp\/data-science-portfolio-examples\/","title":{"rendered":"24 Best Data Science Portfolio Examples 2026 (Analysts & ML Engineers)"},"content":{"rendered":"\n

The best data science portfolios show a problem, the data, the method and a measured result, then link the code and a live dashboard or app so anyone can check the work.<\/strong> These 24 come from data analysts, data scientists, machine learning engineers, data engineers and data visualization designers, from new graduates and career changers to senior practitioners. If you build software rather than analyses, see our developer portfolio examples<\/a>.<\/p>\n\n\n\n

We tested 113 web addresses from published roundups, community lists and our own searches, then opened project pages, repositories and live demos on every shortlisted site, and identified each platform from the site’s code rather than guessing. Every entry is the person’s own website, not a GitHub, Kaggle or LinkedIn profile, and each showed recent work when we checked. Sites that had gone quiet, redirected elsewhere or turned into course-sales pages were dropped.<\/p>\n\n\n\n

Best Data Science Portfolio Examples<\/h2>\n\n\n\n

1. Aditya Ranjan<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Custom HTML (GitHub Pages)<\/p>\n\n\n\n

\"Aditya<\/figure>\n\n\n\n

Aditya Ranjan, a data analyst in Noida, India, opens his portfolio with one result from each of his three projects, such as an 11.61% default rate across 255,347 loans that rises to 19.71% in the riskiest segment. Each project then states the question, the data and the method.<\/p>\n\n\n\n

His retention project builds a three-layer Snowflake warehouse and classifies 4,725 products with a locally hosted LLM, so no transaction data left his machine. Projects link live Tableau dashboards or repositories with methodology notes, and the skills list names only tools a repository demonstrates.<\/p>\n\n\n\n

What stands out:<\/strong> Put one measured result from each project on the first screen.<\/p>\n\n\n\n

2. Harrison Jansma<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Custom HTML (GitHub Pages)<\/p>\n\n\n\n

\"Harrison<\/figure>\n\n\n\n

Harrison Jansma leads generative AI for Capital One’s contact centers, and his portfolio solves a problem most ML engineers share: the best work is proprietary. So he rebuilt the shape of a real-time call copilot in public, on Deepgram and OpenAI, and anyone can try the live demo or read the code.<\/p>\n\n\n\n

The write-up admits it isn’t a machine-learning model he trained. His Projects page lists ten systems from prototype to production, filterable by type, and his About page traces the path from a math and business degree at Baylor to managing an ML team.<\/p>\n\n\n\n

What stands out:<\/strong> Rebuild confidential work as a public demo you can share.<\/p>\n\n\n\n

3. Nicola Rennie<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Quarto (GitHub Pages)<\/p>\n\n\n\n

\"Nicola<\/figure>\n\n\n\n

Nicola Rennie is a data visualisation specialist with a background in statistics and data science, and her Quarto site shows how to organise a varied body of work. The home page is just four buttons, and the Projects page files 16 projects under filters such as Dashboards & Apps, R, Python, Research and Teaching.<\/p>\n\n\n\n

Books, colour palette apps, R packages and the Best Practices for Data Visualisation guidance she co-authored for the Royal Statistical Society all count as projects. Separate Talks and Blog sections keep the project index short enough to scan.<\/p>\n\n\n\n

What stands out:<\/strong> Let visitors filter projects by tool and type.<\/p>\n\n\n\n

4. Marcos Saade<\/a><\/h3>\n\n\n\n

Built with<\/strong>: React (GitHub Pages)<\/p>\n\n\n\n

\"Marcos<\/figure>\n\n\n\n

Marcos Saade is a machine learning engineer whose case studies read like engineering reports. Each opens with his role, the stack, the context and the outcome, so a reader knows at a glance that the orthopedic scanner was freelance work now serving 4,000+ patients a year, while the smart cooler began as a Qualcomm internship demo.<\/p>\n\n\n\n

His first screen leads with three measured results, including a clinic workflow cut from around 10 minutes per patient to under 3 and an 8x edge inference speedup. Each smaller case study links its GitHub repository.<\/p>\n\n\n\n

What stands out:<\/strong> Open every case study with role, stack, context and outcome.<\/p>\n\n\n\n

5. Lorenzo Miguel Bo<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Custom HTML (GitHub Pages)<\/p>\n\n\n\n

\"Lorenzo<\/figure>\n\n\n\n

Lorenzo Miguel Bo moved into data analytics from insurance operations, and his project pages show dashboards working rather than as screenshots. Each embeds the interactive Power BI report and a video walkthrough under a summary and a problem statement.<\/p>\n\n\n\n

His Philippine CPI dashboard asks where inflation actually comes from, using official PSA data across five report pages and regression modeling on Brent crude oil prices. He is also clear about sources: the flood control dashboard is labelled as a synthetic dataset built for a Power BI training workshop.<\/p>\n\n\n\n

What stands out:<\/strong> Embed the live dashboard, then explain the question it answers.<\/p>\n\n\n\n

6. Mehdi Ouazza<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Next.js<\/p>\n\n\n\n

\"Mehdi<\/figure>\n\n\n\n

Mehdi Ouazza is a data engineer and developer advocate at MotherDuck, based in Brussels, and his Next.js site works as a hub for everything he makes. The home page counts his blog posts, videos and GitHub repositories, and the Projects page gives each tool a description, tech tags and a GitHub star count.<\/p>\n\n\n\n

His About page is a compact career history, from Spark clusters at AXA to Klarna, Back Market and Trade Republic, with results such as leading a data warehouse migration from AWS to GCP. Recruiters can ask for a one-page PDF version.<\/p>\n\n\n\n

What stands out:<\/strong> Show traction, such as GitHub stars, next to each project.<\/p>\n\n\n\n

7. Nikita Kumari<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Custom HTML with Tailwind CSS (GitHub Pages)<\/p>\n\n\n\n

\"Nikita<\/figure>\n\n\n\n

Nikita Kumari moved from biology into a master’s in data science, and every project card carries three links: Live Demo, GitHub and Project Story. Her fraud detection project runs from a PySpark pipeline on Databricks to a FastAPI service on Render, and reports 78% recall at 88% precision.<\/p>\n\n\n\n

The repository explains why accuracy is missing: with a 0.17% fraud rate, predicting no fraud for everything would score 99.8%. Her first screen also answers a recruiter’s practical questions, stating that she is willing to relocate and authorized to work in the US for any employer.<\/p>\n\n\n\n

What stands out:<\/strong> Report the metric that fits the problem, and say why.<\/p>\n\n\n\n

8. Kerry Kolosko<\/a><\/h3>\n\n\n\n

Built with<\/strong>: WordPress<\/a><\/p>\n\n\n\n

\"Kerry<\/figure>\n\n\n\n

Kerry Kolosko is a social scientist working in business intelligence, and her site shows that an analyst can build a portfolio from report design alone. The Visualisations page is a gallery of Power BI reports labelled by technique, such as Deneb, Charticulator or HTML Content, from Bob Ross paintings to Adelaide rainfall.<\/p>\n\n\n\n

A Templates page gives away Deneb and PlotlyJS chart templates and theme files, and her blog covers details other Power BI users miss, drawn from almost a decade of building reports. Her About page lists her talks, from Melbourne to Oslo.<\/p>\n\n\n\n

What stands out:<\/strong> Label each dashboard with the technique behind it.<\/p>\n\n\n\n

9. Shaun Lin<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Next.js (GitHub Pages)<\/p>\n\n\n\n

\"Shaun<\/figure>\n\n\n\n

Shaun Lin is a data engineer with an analytics master’s from Georgia Tech, and his case studies draw the pipeline stage by stage. The retail supply chain project runs from data cleaning to a star schema of 7 tables, loading into SQLite, DuckDB or MySQL, an automated integrity audit and analytical SQL.<\/p>\n\n\n\n

Every case study opens with metric tiles and a problem statement, and the retail pipeline has a live demo alongside its repository. A Timeline page replaces the CV, with results such as cutting average dbt model build time by 20% at American Student Assistance.<\/p>\n\n\n\n

What stands out:<\/strong> Draw the pipeline as numbered stages.<\/p>\n\n\n\n

10. Eugene Yan<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Static site (GitHub Pages)<\/p>\n\n\n\n

\"Eugene<\/figure>\n\n\n\n

Eugene Yan is an ML engineer at Anthropic, previously a principal applied scientist at Amazon, and his site keeps side projects on their own Prototyping page. Each gets one line, a date and links to the code and a demo, from AlignEval, an app for building and evaluating an LLM-evaluator, to a recommender built on semantic IDs.<\/p>\n\n\n\n

His About page quantifies earlier work at Alibaba and Lazada, such as product ranking that lifted conversion by 5 to 20% and classification that cut costs by 90%. Writing and talks have their own pages.<\/p>\n\n\n\n

What stands out:<\/strong> Give side projects their own page with code and demo links.<\/p>\n\n\n\n

11. Yu Dong<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Jekyll, Minimal Mistakes theme (GitHub Pages)<\/p>\n\n\n\n

\"Yu<\/figure>\n\n\n\n

Yu Dong is a data scientist with 8 years in product growth analytics, and this portfolio rests on a habit: a new Tableau visualization every week, about 400 so far. The Data Viz page indexes them by year, each week titled with its subject, such as Salary Needed to Afford Rent or Caltrain Weekday Ridership.<\/p>\n\n\n\n

Articles on using AI across the data science workflow and bi-monthly reading notes sit alongside, all on the open-source Minimal Mistakes theme for Jekyll. It shows that consistency can be the strongest signal a data portfolio sends.<\/p>\n\n\n\n

What stands out:<\/strong> Publish on a schedule and index the archive.<\/p>\n\n\n\n

12. Maggie Wolff<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Jekyll, AcademicPages theme (GitHub Pages)<\/p>\n\n\n\n

\"Maggie<\/figure>\n\n\n\n

Maggie Wolff moved from marketing and PR into data science, and her home page says so in its first line. She says she enjoys translating complex findings for stakeholders as much as doing the analysis, and her CV page traces the path from hospital PR to data scientist at Box.<\/p>\n\n\n\n

Her Portfolio page is short, with four machine learning projects, including a running clothes recommender built on her own data. Talks and a podcast show her community work as a Women in Data Science Chicago ambassador. The site runs on the AcademicPages theme for Jekyll.<\/p>\n\n\n\n

What stands out:<\/strong> Turn a previous career into a selling point.<\/p>\n\n\n\n

13. Nadieh Bremer<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Hugo<\/p>\n\n\n\n

\"Nadieh<\/figure>\n\n\n\n

Nadieh Bremer is an award-winning data visualisation designer, and her portfolio can be filtered by chart type, from beeswarm and network to radial and treemap, as well as by media and industry. Client tiles on the home page, from Google Trends to UNESCO, set the level before a project opens.<\/p>\n\n\n\n

Each project page lists the client, awards, categories and tools, such as d3.js, Python and R. Searching for Birds, built for Google Trends from search data on nearly 700 species, won a People’s Voice Webby Award.<\/p>\n\n\n\n

What stands out:<\/strong> Tag every project with the tools used.<\/p>\n\n\n\n

14. Rich Pauloo<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Hugo<\/p>\n\n\n\n

\"Rich<\/figure>\n\n\n\n

Rich Pauloo is a data scientist who works on water, climate and infrastructure data, and his portfolio publishes the working behind a result. His New York subway accessibility dashboard comes with a companion notebook that shows the data transformations, the exploratory analysis and the results, plus the GitHub repository.<\/p>\n\n\n\n

The home page is a garden of projects and publications, each tagged and dated, from interpretable random forests to groundwater papers. His About page explains the thread running through it all: making invisible systems visible, from underground water to energy infrastructure.<\/p>\n\n\n\n

What stands out:<\/strong> Publish the notebook behind the dashboard.<\/p>\n\n\n\n

15. Deshraj Jogiya<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Custom, Python and Jinja2 (GitHub Pages)<\/p>\n\n\n\n

\"Deshraj<\/figure>\n\n\n\n

Deshraj Jogiya is a new data engineering graduate from Arizona State University, and his portfolio is built for recruiters in a hurry. A Recruiter Tour button walks through the page, an AI assistant answers questions about his experience, and the r\u00e9sum\u00e9 opens in the page with a Print \/ Save PDF option.<\/p>\n\n\n\n

Four tiles sit on the first screen, including 11+ production pipelines and 2300+ Git commits. Below, each role lists quantified results, and a filterable list links to his repositories, from a dbt warehouse to a job-market lakehouse.<\/p>\n\n\n\n

What stands out:<\/strong> Make the portfolio easy for a recruiter to skim.<\/p>\n\n\n\n

16. Hamel Husain<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Quarto (GitHub Pages)<\/p>\n\n\n\n

\"Hamel<\/figure>\n\n\n\n

Hamel Husain is a machine learning engineer with 20+ years of experience, including Airbnb and GitHub, and his Quarto site shows open-source work as a portfolio. The Open Source page lists each project with his role: creator of fastpages and core contributor to nbdev, fastcore and ghapi, with links to related posts and talks.<\/p>\n\n\n\n

The home page is a dated table of long-form writing, most of it on evaluating AI products, the subject he now teaches. A Teaching page collects his courses and talks, many with slides and video.<\/p>\n\n\n\n

What stands out:<\/strong> Say what your role was on each open-source project.<\/p>\n\n\n\n

17. Georgios Karamanis<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Blot<\/p>\n\n\n\n

\"Georgios<\/figure>\n\n\n\n

Georgios Karamanis is a data visualization designer, psychiatrist and researcher in Uppsala, Sweden, and his home page is his weekly practice, 200+ TidyTuesday charts and counting. Each post says what data he used, what he calculated and what he found, such as two in three of 1,112 cities losing green space.<\/p>\n\n\n\n

He also flags caveats, noting when the time spans behind the numbers are not all the same, and links the code on GitHub every week. A short About page lists client work for Scientific American and Visual Capitalist. The site runs on Blot.<\/p>\n\n\n\n

What stands out:<\/strong> State the finding and the caveat, then link the code.<\/p>\n\n\n\n

18. Kelly J. Adams<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Wix<\/a><\/p>\n\n\n\n

\"Kelly<\/figure>\n\n\n\n

Kelly Adams studied secondary education before taking the Google Data Analytics Course, and her blog documents the switch month by month until she started as a data analyst. She now works as an analytics engineer, and her Wix portfolio shows two projects from her job.<\/p>\n\n\n\n

Each has a GitHub repository and a plain-language article, and the churn model’s README notes that database names and query structure were changed to protect private information. It is a practical way to show work projects without breaking confidentiality.<\/p>\n\n\n\n

What stands out:<\/strong> Anonymise work projects instead of leaving them out.<\/p>\n\n\n\n

19. Randal S. Olson<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Next.js<\/p>\n\n\n\n

\"Randal<\/figure>\n\n\n\n

Randal S. Olson has worked in AI for 15+ years, and his home page opens with numbered selected work, each labelled by type, such as blog series or open source. TPOT, his AutoML tool for Python, is cited thousands of times.<\/p>\n\n\n\n

His optimal road trip project, which used a genetic algorithm to solve the traveling salesman problem across all 50 U.S. states, links the Jupyter notebook behind it. A r\u00e9sum\u00e9 page and a Work with me page explain what he does now: building independently and taking short advisory engagements.<\/p>\n\n\n\n

What stands out:<\/strong> Number your best work and label what kind it is.<\/p>\n\n\n\n

20. Jadey Ryan<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Quarto<\/p>\n\n\n\n

\"Jadey<\/figure>\n\n\n\n

Jadey Ryan is the data scientist for the natural resources section of the Washington State Department of Agriculture, and her portfolio shows that tools built at work count. Projects include a Shiny app that generates soil health reports, R packages for her team’s branding and reporting, and a soil health roadmap built with Python and ArcGIS.<\/p>\n\n\n\n

Each project is tagged by type and dated, and filters separate R packages, Shiny apps and Quarto work. Her first personal project, built to learn web scraping, stays in the list, and a full CV page sits in the menu.<\/p>\n\n\n\n

What stands out:<\/strong> Count the tools you built at work.<\/p>\n\n\n\n

21. Simon Sp\u00e4ti<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Hugo<\/p>\n\n\n\n

\"Simon<\/figure>\n\n\n\n

Simon Sp\u00e4ti is a data engineer, author and educator, and his site makes writing the portfolio. Blog posts, a second brain of 1,000+ notes and chapters of his book on patterns of data engineering are linked in one interactive graph that visitors can drag and explore.<\/p>\n\n\n\n

The Data Engineering Vault works as a glossary of the field, and the first chapters of the book are free to read. For a data engineer whose work lives in internal systems, that much public writing shows how he thinks.<\/p>\n\n\n\n

What stands out:<\/strong> Link your writing so readers can follow a thread.<\/p>\n\n\n\n

22. Tyler Richards<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Custom HTML with Streamlit apps (GitHub Pages)<\/p>\n\n\n\n

\"Tyler<\/figure>\n\n\n\n

Tyler Richards is a data scientist who liked Streamlit enough to write a book about it and join the company behind it, and his site is a single page of projects. Several are live Streamlit apps, such as a Goodreads analytics app, so visitors can try the work instead of reading about it.<\/p>\n\n\n\n

Older write-ups remain, including an Elo ranking system for ping pong built from scratch in Python. The page states when it was last updated, a simple sign that it is maintained.<\/p>\n\n\n\n

What stands out:<\/strong> Ship small projects as apps people can use.<\/p>\n\n\n\n

23. Samuel Tseng<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Jekyll, Chirpy theme (GitHub Pages)<\/p>\n\n\n\n

\"Samuel<\/figure>\n\n\n\n

Samuel Tseng is a data engineer whose blog doubles as his project portfolio. Posts such as Building a Production-Ready Data Pipeline with Airflow and dbt explain the challenge, the architecture and the design decisions, then link the repository with the code.<\/p>\n\n\n\n

Categories separate data engineering, learning journeys and career reflections, including a year-end post on landing a new job. The site runs on the Chirpy theme for Jekyll and GitHub Pages, and his About page links a CV and LinkedIn.<\/p>\n\n\n\n

What stands out:<\/strong> Write up each project as a post and link the repository.<\/p>\n\n\n\n

24. Beatriz Milz<\/a><\/h3>\n\n\n\n

Built with<\/strong>: Quarto<\/p>\n\n\n\n

\"Beatriz<\/figure>\n\n\n\n

Beatriz Milz is an environmental data science researcher in Brazil and a certified tidyverse instructor, and her Quarto site is published in English, Portuguese and Spanish. The home page routes visitors to the about page and the blog in each language.<\/p>\n\n\n\n

Her projects are open-data R packages, such as one with Brazilian origin-destination survey data, plus community work, including coordinating the Portuguese translation of R for Data Science. Her About page notes that she is a GitHub Star.<\/p>\n\n\n\n

What stands out:<\/strong> Publish in the languages your audience reads.<\/p>\n\n\n\n

What Makes a Great Data Science Portfolio?<\/h2>\n\n\n\n

Lead with the business question.<\/strong> A model or dashboard means little until the reader knows what it was for. Aditya Ranjan’s process starts with the business question and each of his projects opens with one, Lorenzo Miguel Bo gives every dashboard a problem statement, and Shaun Lin begins each case study with the problem and the business objective before any architecture.<\/p>\n\n\n\n

\"Shaun
Shaun Lin opens each case study with the problem and the business objective before the pipeline.<\/figcaption><\/figure>\n\n\n\n

Link the code and a live demo or dashboard.<\/strong> The repository proves you wrote it; the demo proves it works. Nikita Kumari gives every project a Live Demo, GitHub and Project Story link, Harrison Jansma rebuilt confidential work as a public demo with its code, Eugene Yan lists code and demo links for each prototype, and Tyler Richards ships projects as Streamlit apps.<\/p>\n\n\n\n

\"Nikita
Nikita Kumari’s project cards each link a live demo, the GitHub repository and a project story.<\/figcaption><\/figure>\n\n\n\n

Quantify the result, with the right metric.<\/strong> Marcos Saade reports a clinic workflow cut from around ten minutes per patient to under three, Aditya Ranjan shows that one customer segment drives most of the revenue, and Shaun Lin cut dbt build times. Nikita Kumari explains why she reports precision and recall instead of accuracy, which is useless when fraud is rare.<\/p>\n\n\n\n

\"Marcos
Marcos Saade’s case study header gives his role, the stack, the context and the measured outcome.<\/figcaption><\/figure>\n\n\n\n

Write it up, in a post or a notebook.<\/strong> Hiring managers want to see how you think, not only the output. Rich Pauloo publishes the notebook behind his subway dashboard, Samuel Tseng turns each pipeline into a long post with the repository linked, Georgios Karamanis states the finding and the caveats every week, and Simon Sp\u00e4ti links his posts, notes and book chapters in one graph.<\/p>\n\n\n\n

\"Rich
Rich Pauloo publishes the notebook behind his subway accessibility dashboard.<\/figcaption><\/figure>\n\n\n\n

Analysts: show the dashboard itself.<\/strong> A screenshot hides what the report can do. Lorenzo Miguel Bo embeds the interactive Power BI report with a video walkthrough, Aditya Ranjan links his live Tableau Public dashboards, Kerry Kolosko labels every Power BI report by the technique behind it, and Yu Dong indexes a weekly Tableau habit by year.<\/p>\n\n\n\n

\"Lorenzo
Lorenzo Miguel Bo embeds the interactive Power BI report on the project page.<\/figcaption><\/figure>\n\n\n\n

Keep it current.<\/strong> Recent, dated work tells a reader you are still practising. Yu Dong has published a Tableau piece every week for years, Georgios Karamanis posts a TidyTuesday chart each week, and Tyler Richards puts the date of his last update at the top. Many portfolios in published roundups have gone quiet, and one domain listed in a data analytics roundup now leads to a gambling site.<\/p>\n\n\n\n

\"Yu
Yu Dong indexes a weekly Tableau visualization by year.<\/figcaption><\/figure>\n\n\n\n

Keep the r\u00e9sum\u00e9 one click away, and test the link.<\/strong> Aditya Ranjan puts a r\u00e9sum\u00e9 download on his first screen, Harrison Jansma and Lorenzo Miguel Bo put one in the menu or beside the projects, Maggie Wolff keeps her CV as a web page, and Mehdi Ouazza offers a one-page PDF. One CV button on our shortlist returned an error, so click yours after every change.<\/p>\n\n\n\n

\"Maggie
Maggie Wolff keeps her CV as a web page, one click from her home page.<\/figcaption><\/figure>\n\n\n\n

New grads: two or three end-to-end projects beat ten tutorials.<\/strong> Aditya Ranjan shows three projects that each run from raw data to a dashboard, with a table comparing the skills each one used. Nikita Kumari’s fraud project goes from a Databricks pipeline to a deployed API, and Lorenzo Miguel Bo labels course and synthetic data honestly. Depth, and a clear account of what you did, beats a long list of copied notebooks.<\/p>\n\n\n\n

\"Aditya
Aditya Ranjan compares the skills each of his three projects used.<\/figcaption><\/figure>\n\n\n\n

Frequently Asked Questions<\/h2>\n\n\n\n