After the dashboard launch we completed a lot of updates in April to make your user experience better.
Various metadata editor enhancements; histogram views vary based on the data type; dataset metrics change inline when data type is changed; fixed box plots on anomaly detection; time-series visualizations now respect time-zones and scale correctly; various bug squashing;
Fixed various issues including the API preventing vocabularies from loading when requested was filtered; paging info in feature importance properly included in responses; performance improvements calculating dataset statistics; improved forecasting and impact analysis session prioritization and more…
Bug fixes: Stop recalculating column statistics when no data has changed; Stop warning about aggregation strategy issues on non-time-series sessions since they have no implications; Add constraints on dataset during pre-processing to not allow engineered features to get too large to process
New metadata editor and histograms visualization; dataset metrics vary based on data-type selected and highlight missing data and data conversion error counts; import larger CSV/JSON files from AWS S3 and Azure Blob storage; box and whisker plot visualization for anomaly detection session results;