The Story

The Humor Index started the way most projects do: lying in bed, unable to sleep, with an idea that wouldn't shut up.

The question was simple — which sitcom is actually the funniest? Not according to critics, not by ratings, not by how loud the laugh track is. By the jokes themselves. Every single one, scored on craft, impact, and density.

I'd had this idea rattling around for years. Then one night it clicked that AI had finally gotten good enough to do it. So I built a pipeline that reads every transcript, detects every joke, and scores each one across multiple dimensions. No opinions, no fan votes — just structured analysis of what makes comedy work.

This is my first project like this. It turns out quantifying humor is harder than it sounds — but the data is genuinely fascinating.

Every episode gets fed through a two-phase AI analysis pipeline. Phase one detects every comedic moment in the transcript — dialogue jokes, physical comedy cues, reaction beats, cringe sequences, callbacks, the works. Phase two scores each joke on craft (originality, structure, character fit, economy, earned-vs-cheap) and impact (audience reaction, quotability, rewatch value).

Those joke-level scores get aggregated into episode scores, season scores, and show-level rankings. The full math is on the Methodology page.

I'm Sam. By day, I work in education and business leadership — from student advisor at Kaplan Test Prep all the way to VP of Growth at Triad. I've spent my career building teams, scaling organizations, and figuring out how to make complex things accessible.

The Humor Index is a different kind of project for me. It's not my industry, it's not my day job — it's the thing I built because I couldn't stop thinking about it. Turns out the skills transfer: breaking down subjective things into measurable components is what I've been doing my whole career, just usually with test scores instead of joke scores.

I live in Orlando, Florida with my wife Franchesca and our three kids — who are, for the record, not yet old enough to appreciate Seinfeld.

LinkedIn

Right now we've fully analyzed The Office and are working through Seinfeld. The goal is complete coverage of 11 major sitcoms — every episode of Friends, Arrested Development, Parks and Rec, 30 Rock, Brooklyn Nine-Nine, It's Always Sunny, Schitt's Creek, The Big Bang Theory, and Two and a Half Men.

After that? More shows based on your votes. Character-level deep dives. Cross-show joke type analysis. And whatever else the data reveals.