Data scientist Anna Smith discusses the variety of ways that Bitly is using Salience’s concept topics for automatic text categorization. Anna discusses the variety of different projects Bitly has teamed up with others to explore, including national attention to the Trayvon Martin case, how many people are clicking on celebrity urls, and most importantly, to find the best kittens on the internet.
In the talk Anna explains how Bitly feeds content from the website of the urls they shorten, and from the websites people are on before or after the click a Bitly link, and feeds them into Salience, where they are categorized. Those categorizations are then used in various ways, including Bitly’s own search engine, which powers their Realtime content discovery engines, and audience analysis.
Anna Smith is currently an analytics engineer at Rent the Runway where she works on solving multiple and various problems from setting up A/B analysis frameworks to validating machine learning implementations in products. Previously a data scientist at Bitly, where she was wooed away from a physics Ph.D program, she is not shy about dealing with algorithms and large data sets. In her spare time you can find her catching up on the latest celebrity gossip or creating her own GIFs. FYI, she is adamant that it is pronounced like “gift”.