Graph Gurus Episode 16: Hyper-Personalized Recommendation Engine Powered by a Native Parallel Graph
wednesday july 24, 2019 at 9:30 AM PST
This is the 16th episode that continues TigerGraph's Graph Guru series, a free educational webinar series for developers and data scientists.
Recommendation engines have been around since mid 1990s. Every time we shop on Amazon.com or another eCommerce site, we see recommendations such as “People who bought this item also bought” and “Items often bought together”. These engines are based on collaborative filtering, a method used to create recommendation based on users who have similar purchases in the past. Innovators such as Wish.com and Kickdynamic are building the next generation of recommendation engines that go beyond basic collaborative filtering to factor in demographics, product features, location, and user preferences and their journey based on browsing search and purchase history. The results speak for themselves - Wish.com has grown from a startup in 2015 to the eCommerce juggernaut with billions in annual revenue powered by their hyper personalized recommendation engine.
In this Graph Gurus episode, we will use the TigerGraph Developer Edition to build a hyper-personalized recommendation engine to create dynamic offers in real-time that result in higher click-through and eCommerce order value. In this episode, we will:
- Explain the new approach for building recommendation engines and review multiple use cases where it can be applied for sectors including retail, eCommerce, telecom, banking and financial services.
- Present a graph solution to model, visualize and optimize a hyper-personalized recommendation engine using TigerGraph GraphStudio and GSQL.
- Review a test dataset and GSQL queries for the solution
If you can't attend the live webinar, sign up for the recording.
Richard Henderson, Solution Architect, TigerGraph