hovanter / cs448b-finalproj Goto Github PK

If < some data type > ==> then use < some UI element >

Right now, the user can specify all columns as interactive.

If we do this… I wonder… if the map with dots is the primary visualization, couldn’t there still be interaction? Like you can hover on the dots?

currently just hard coding what children I am going to for the demo

If its greater then one make the values non decimal would be smart I think

Given a Vega-Lite visualization, figure out

(1) Extract all the data types that are in it. After snooping around a bit this afternoon, I figured out you can do something like

d3.selectAll("<some css class here>").data()[0].datum()

.data() returns an array of all the data bound to all the elements with the given CSS class. Given an item in that array, you can call .datum() to get a specific data point.

(2) Create the appropriate GUI for filtering each data type (todo with @hovanter and the decision tree issue). Maybe just some simple Javascript/jQuery to dynamically generate some new DOM nodes containing the GUI elements and stuff it at the end of the body tag.

Have a primary visualization and a secondary table that displays information from the data that is not shown in the visualization. For example, this might be summary statistics like # of data points, average, median, variance, etc.

Filters apply over both primary and the table. If we have time, we can turn the table into a secondary visualization.

Take the user CSV/JSON, and generate a Vega-Lite spec, and show the resulting viz on the output page.