This is an excellent post. I love the instrumentation you discussed for measuring and improving front-end performance. I also love the use of a graph db for modeling the fast lookups. Three questions :)
1) Have you published anything on the internal front-end performance instrumentation? I’d love to learn more about what this look likes and how it works.
2) What graph db did you select? This would also make a great follow up post on the data model and how it works at a deeper level.
3) It sounds like the history is stored and scores calculated locally for performance reasons… is it synced between platforms/devices to make the user experience consistent?
Thanks for such a great post!