Michael Nocito.
I find the signal in messy data and explain it clearly — then build the tools to do it again, faster. Python, SQL, Tableau, and pandas, applied to real datasets and shipped as working software.
10+ years migrating data and documents across enterprise systems, now focused on data analysis with Python, SQL, pandas, and Tableau (certification in progress). A research and training background taught me the core of the job: find the signal in messy data, explain it clearly, and make it easy to act on. Below is a working portfolio — real analysis, learning tools, and the utilities I built to do the work. Everything is live and linked.
SQL analysis of a real ~125,000-game Steam dataset to surface 166 highly rated games almost nobody has played, including catching and fixing a broken source file before the numbers could be trusted. A Part 2 expansion joins ~41 million player reviews to see which of those gems people actually keep playing.
SQL analysis of IMDb's full 12.6-million-title dataset to surface 581 movies rated 8.0+ that almost nobody has watched, including catching an import that silently dropped 256,000 rows before the numbers could be trusted. Every threshold measured from the data, every filter's cost counted. Companion to Steam Hidden Gems.
SQL analysis joining 354,500 weeks of Billboard Hot 100 history (1958–today) to 78,335 Last.fm tracks to surface 500 songs by Top 40 artists that listeners love but that never made mainstream airplay. Required real entity resolution — matching messy artist credits and song titles across two sources — plus a Python script pulling listener data from the Last.fm API. Third entry in the Hidden Gems series.
A complete browser-based prep suite for breaking into data analytics, sequenced in the order an analyst actually learns: Excel → SQL → Python → Tableau → Stats → Power BI → Interview, plus tool-agnostic Chart Literacy and Forecasting kits, a week-1 job simulator, a cross-subject final exam, and certification prep for Tableau, Excel, and Power BI. Read-it-aloud code breakdowns and live in-browser practice environments. No install, no login, no telemetry.
Ten-minute warm-up games for the prep kits. Coach Srbina, a sales rep finding her analyst instincts, through 14 bite-size tasks per skill — type real Excel formulas (=SUM to XLOOKUP), then real SQL queries that run against an in-browser database, with instant feedback and a story that carries from game to game. One screen, one task, zero clutter.
“Getting it wrong gets you there faster.” A day-in-the-life data-analyst job sim with Excel and Tableau tracks: train with a mentor, then de-curse the messy files inherited from the last person who had your job — learning good habits by fixing bad ones. One reusable engine, two skill tracks.
A roguelike SQL tower-defense game running real SQLite in the browser. You hold off each wave by writing queries — the better the query, the better the defense.
Generate realistic fictional test documents and data — PDF, Word, HTML, Excel — for QA, ETL, and data-engineering testing. Fully offline, no real customer data, with a free Windows installer.
A beginner-friendly Python project that profiles a messy spreadsheet and flags data-quality issues — row and column counts, plus missing values per column — in a plain audit report. Built with guided, plain-English comments to learn data cleaning from the ground up.
A calm typing trainer for the symbols data work actually runs on — brackets, braces, operators, quotes. A biodiversity garden grows as you practice, so the reps stay low-pressure.
The step nobody explains: turn a CSV file into a real SQL database on your own computer with free tools, no server. Create, import, verify, and fix the two import problems that bite everyone once. Written for complete beginners.
A file too big to open in Excel? What "large" really means, why you move it into a database, and the real quirks that trip everyone up: the zipped download, slow imports, indexing, sampling, and data types that quietly break a filter. Uses a real 2 GB, 41-million-row example.
The small, repeatable core of Git an analyst actually needs: the everyday add-commit-push loop, starting a repo, publishing a project as a live site with GitHub Pages, what not to commit, and the five errors everyone hits once.
The full, growing library: the how-tos above plus concept guides on the SQL and analysis ideas that trip up beginners, COUNT, CASE, exploratory data analysis, defining metrics, choosing thresholds, and documenting a dataset's limits.