Theodore Speaks about TED Talks

Learn to communicate better from the most persuasive speakers in the world with tools and insights from the Theodore Speaks web app.

The purpose of this app is to understand (and emulate) the content and linguistic style of highly persuasive talks.

The app uses natural language processing techniques and the Linguistic Inquiry and Word Count (LIWC) module to analyze the transcripts of 2600+ TED Talks. The embedded models include decision trees, random forest regressors, and linear regression to find text categories with statistically significant relationships to user's ratings and to the number of times the talk has been viewed.

Choose the insights you want to see in the dashboards.

Select between Persuasive, Inspiring, Views and Moral Signature.

Recently Added Talks

No bookmarks yet.

Learn How to Give More Persuasive, Inspiring or Engaging Presentations

Ask Theodore about TED. Go ahead and poke and click the charts below to see what they do.

Predict Persuasiveness of Text

The purpose of this tool is to predict the persuasiveness of a text. It uses an embedded Multinomial Naive Bayes classifier. Put the text of a speech in the field below, and have it predict whether the text is persuasive.



Find Similar TED Talk Speaker

The purpose of this tool is to find a talk that is similar in linguistic style and content to a speaker you specify. It uses an embedded Euclidean distance calculation to determine similarity. Put a TED Talk speaker's name in the field below, and the tool will find a similar talk.


Talk with Most Similar Linguistic Style and Content