An exercise in data science and text. Not a page about what Rory Sutherland said — a page about what the public record of his talks, articles and podcast appearances looks like once you stop reading it as prose and start reading it as data.
Built entirely from titles, dates and venues that are already public — TED's speaker index, publisher pages, show notes, article archives. No transcript, talk audio, or book text is reproduced anywhere on this page. Where you want his actual words, every tab links straight to the people who hold them.
Rory Sutherland is a working example of a specific problem: a public figure whose ideas exist almost entirely as scattered records — a TED talk here, a Spectator column there, ninety-odd podcast credits spread across a dozen platforms. This installation takes that scatter and asks a data-science question of it: what does the shape of a public output look like, independent of its content?
Three views, three techniques. A word cloud of terms that recur across the public titles — a classic, slightly blunt way of turning text into an image. A frequency chart and timeline — the same underlying counts, read as numbers instead of typography, plus a year-by-year count of dated public appearances, which turns a career into a bar chart. And word art — the same vocabulary, made decorative rather than analytical, because a data exercise that never admits it's also an aesthetic choice is lying about half of what it's doing.
The catalogue is the bibliography underneath all three: real titles, real dates, real links to TED, Penguin, The Spectator, Behavioral Scientist and the rest. The sources tab says plainly what this page is and is not.
Built from the words in ~90 publicly catalogued talk, podcast-episode and article titles about or by Rory Sutherland — not from the talks or articles themselves. Size is frequency across those titles. The name itself is excluded; a word cloud of "Rory Sutherland" repeated ninety times would not tell you anything.
Methodology note: title words only, hand-tallied from the research index below, stopwords removed. This is a study of headline vocabulary, not of argument — a title is a marketing decision made by an editor, not a transcript.
Top terms by raw count across the catalogued titles, and a year-by-year count of dated public appearances (talks, podcast episodes, articles) from the same index. Undated and future-listed items are excluded from the timeline; it undercounts rather than guesses.
Reads as a rough proxy for media presence, not popularity or quality — a busier year of podcast bookings looks identical to a busier year of good ideas on this chart, and the chart cannot tell you which.
The word cloud's terms, set adrift. No new information here — this tab exists to be honest that "data as art" is doing aesthetic work as much as analytical work, and to let the two sit next to each other rather than pretending the chart tab is the objective one and this tab is the fun one.
A representative sample of the public record this page draws its counts from — books, TED talks, and a spread of podcasts and articles, each linking to the original, official page. Not exhaustive: treat TED's own speaker page and the publisher archives linked here as the living, current index.