Forking memories sounds like a Black Mirror episode, but it’s totally chill for analytics, no dystopian futures — we promise. In fact, it introduces interaction patterns and opportunities for AI beyond those found in forking code or creating a copy of a viz or doc.
Instead of just a copy — where intuition is lost once your changes are made — forking memories allows you (and HyperArc) to remember how you got there and the critical inspiration behind your own insights.
Aggregated across HyperArc, forks in your analytics memory are like the hyperlinks that enabled Google to automate search with PageRank and dominate curated portals like Yahoo.
Lets find out “which schools have the highest replacement cost per supervisor” given SF school’s recent budget issues. I’ve played around with the topic before, so we’re able to pull up relevant memories and use several of them to give us a summary. However, the top memory is not fine grained enough and is not broken down by school and we want to dig deeper.
The second memory allows us to do this and we’re able to navigate to it specifically being just a moment in a larger exploration and not the final visualization saved.
To dig deeper with our own customizations, we’re able to fork it to a new query and seed our new HyperGraph. Not only do we get a copy of the visualization, we retain the walk and siblings to that memory, the intuitions and references critical for building the “page rank” for analytics.
Now if we ask the same question, we see the most relevant memory boosted to the top which also enables a much more succinct answer.
.
Share this post