Visualizing High-Dimensional Product Spaces: An End-to-End Approach

Agenda

6:30 – 6:40 Intro
6:40 – 7:40 Presentation on Visualizing High-Dimensional Product Spaces
7:40 – 7:55 Q&A
7:55 – 8:00 Adjourn

Abstract

How do you search for products? In this meeting, Yaniv will report his latest efforts to help readers find their next book. Using deep-learning, he's created a map of all the books that encodes both genre and popularity. The project involves scraping data with Scrapy, preprocessing with SQL, modeling with Tensorflow, visualization in Javascript, and deployment with Flask on AWS. This project is a set of open-source tools that together cover all you need to go from a data project idea to a sharable proof-of-concept. We'll get a short intro to each tool and see how it all connects into one end-to-end workflow.

Bio

Yaniv Ben-Ami is an Assistant Professor of Economics at Carleton College where he teaches Finance, Macro, and Econometrics. He's the founder of the Twin Cities Deep-Learning Study Group, a weekly meetup for deep-learning enthusiasts. His research focuses on the visualization of high-dimensional data through the use of non-linear dimension reduction techniques. Some of his previous projects include creating visualizations for the American Time Use Survey (ATUS) and daily Covid-19 cases. Yale Ph.D. in Economics, Tel-Aviv University M.A. in Economics and B.Sc. in Computer Science and Biology; self-identifies as a programmer.