[PERF] Barcode Plot - Avoid Rendering Duplicate Lines for the Same Value

Issue #76 resolved
Daniel Marsh-Patrick repo owner created an issue

Looking at introducing #75, if we have a large data set with a low cardinality of measure values but a high cardinality of samples (e.g. movie ratings out of 5 data set where there are 11 possible values but 26M individual samples, we're rendering waaaaaay more SVG lines than we need to - a maximum of 11 is better than millions.

We should consider how we can represent the number of samples against an individual data point and I'll create a different ticket to track that one.

Official response

  • Daniel Marsh-Patrick reporter

    We'll add an attribute for aggregation of data points to each category. For now we'll just keep the count but we'll use an IStatistics in case we need anything else. This can then be used to calcluate the tooltips in #77. If plotting the barcode, we'll switch out for aggregated data points rather than raw data points.

Comments (6)

  1. Daniel Marsh-Patrick reporter

    We'll add an attribute for aggregation of data points to each category. For now we'll just keep the count but we'll use an IStatistics in case we need anything else. This can then be used to calcluate the tooltips in #77. If plotting the barcode, we'll switch out for aggregated data points rather than raw data points.

  2. Log in to comment