We need to develop AI text-processing applications called Content Synthesizer in Wolfram Mathematica to
-
Server custom synthesize Observable notebooks per certain input form the user, in particular I am thinking about a Free Form interface e.g. “Generate a notebook for Bubble map with three sliders for t and x and y with data source at xyxt …”
-
bind the data source to the notebook if there is a match to the variables
-
Server might synthesize 100s or 1000s of such Observable notebooks that go into some indexed dashboard for a corporate use
We already do exactly that in Mathematica for SCADA DATA acquisition for large Wind Turbine farms e.g.
https://my.compclassnotes.com/guest/5337/example/embed
These are produced in large servers computing predict functions or classifiers used for alarms to notify management for near future failure or required maintenance.
These we call DataNotes, are then pushed into an indexed dashboard via mySQL.
While Mathematica does output in SVG, it lacks luster luster compared to certain Observable applications.
We like to find out if we can extract the Javascipt textual code into our AI text processing templates to develop Observable notebook synthesizers then from our server automatically upload the notebooks into Observable hosting servers.
An architectural note: Observable notebook concept, as opposed to the older just D3 libraries, allows our AI applications to construct notebook templates which are far easier to model in Mathematica than loose unstructured code.
Dara