Chanin Nantasenamat Portfolio
Developer Advocate
Ex-bioinformatics professor turned developer advocate ("Data Professor") whose GitHub (290 repos) is the runnable half of a 222K-subscriber YouTube channel: R/Python notebooks spanning cheminformatics and applied ML, backed by what his README counts as >160 peer-reviewed research and review articles.
Location
Didn't say. We're respecting the mystery.
Companies
Skills
Notable
- Data Professor YouTube channel: 222K subscribers, 372 videos
- dataprofessor/code repo: 1.1k stars, 1.4k forks
- streamlit_freecodecamp ('Build 12 Data Apps in Python with Streamlit'): 678 stars, 573 forks
- Ex-Professor of Bioinformatics with >160 research/review articles (per his GitHub README)
What makes it work
A breakdown of the choices that make this portfolio stand out.
Academic publication record backing applied tutorials
His README cites >160 research and review articles "at the interface of machine learning, biology and chemistry," aimed at drug discovery. That background shows up in the repos: the pinned data repo ships acetylcholinesterase bioactivity datasets next to Iris and Boston Housing, a depth check most content-creator data science material lacks.
Consistent, high-volume public teaching output
The README counts >400 videos and ~100 blogs; the channel currently lists 372 public videos, with uploads still landing in 2026. Output sustained that long proves an ability to explain technical work clearly and repeatedly.
Reproducible notebooks as the actual deliverable
Repos are structured as runnable companions to tutorials, so a viewer can execute the exact code shown rather than take a claim on faith.
What Data Scientists can take from this
Specific, actionable tips to apply to your own portfolio — no generic advice.
If you left academia, treat your publication list as portfolio evidence — link it directly.
Make every explainer accompanied by runnable code in the same repo.
Consistency compounds: fewer, well-produced tutorials sustained over years beat a one-time dump.
Pick a domain (bioinformatics, finance, geospatial) to specialize your applied-ML teaching in.
Projects
On the pinned code repo, forks (1.4k) outnumber stars (1.1k): people run Chanin Nantasenamat's tutorials more than they bookmark them. The GitHub profile is the lab bench behind the Data Professor channel, with 290 repositories that are mostly the runnable half of a video.
A compilation of the R and Python code behind Data Professor YouTube tutorials. Folders are split by topic: iris and dhfr for exploratory analysis and classification models, linear-regression, plotting, python-in-r, python project tutorials, shiny web apps and streamlit apps. Viewers can run the exact code shown on screen.
Tutorial code that thousands of people have forked and run is a public test of clarity and correctness. Every notebook is checked by strangers executing it.
Covers both R (including Shiny) and Python in one repo, organized by tutorial topic.
The dhfr folder applies the same EDA-plus-classification workflow to a bioinformatics dataset.
Its repo topics tag scikit-learn, pandas, exploratory-data-analysis and shiny.
Code for a full-length course video, 'Build 12 Data Science Apps with Python and Streamlit - Full Course'. The repo folders are ten numbered apps: a simple stock-price app, a DNA-sequence bioinformatics app, EDA apps for basketball, football, S&P 500 and cryptocurrency data, Iris and penguins classifiers, and regression apps for Boston housing and molecular solubility.
It turns models into apps a stakeholder can click, across finance, sports, biology and chemistry datasets. That is the last-mile skill many notebook-only portfolios skip.
Numbered app folders progress from a simple stock chart to ML regression apps.
Two apps (DNA analysis, molecular solubility) bring his bioinformatics background into beginner material.
The repo title promises 12 apps; ten app folders are visible in the repo root.
Computational Drug Discovery: Acetylcholinesterase Bioactivity ML
A five-notebook, end-to-end bioinformatics project that predicts acetylcholinesterase inhibitor bioactivity. Part 1 collects bioactivity data. Part 2 is exploratory data analysis. Part 3 prepares a descriptor dataset. Part 4 builds a random forest regression model. Part 5 compares regressors. The repo also commits every intermediate CSV, from raw bioactivity data through curated data, pIC50 classes and PubChem fingerprint tables.
Committing each data stage (raw, preprocessed, curated, featurized) makes the pipeline auditable step by step. It is also real domain modeling from a former bioinformatics professor rather than a toy dataset.
Seven numbered CSVs trace the data from raw bioactivity to 2-class and 3-class pIC50 labels with PubChem fingerprints.
Ends with a model-comparison notebook instead of stopping at a single random forest.
Each notebook maps to one stage of a standard QSAR-style workflow.
More Data Scientist portfolios
Ready to build your portfolio?
Follow the Data Scientist roadmap — skills, projects, and timeline to get hired.