Skip to content

v0.1.0

Latest
Latest

Choose a tag to compare

View all tags
@BurakUlver BurakUlver released this 02 Apr 05:47
v0.1.0
28d4774

Synthetic data quality evaluation with interactive HTML reports.

Added

  • SynthEval class — dual-input (DataFrame or ProfileReport), lazy diff computation
  • Dataset overview: comparative row/column counts, missing %, duplicate %, type distribution
  • Column comparison: per-column stat deltas with distribution overlay charts and trend lines
  • Bivariate interactions: scatter plots and box plots with outlier display, dropdown-based column selection
  • Correlation comparison: Original, Synthetic, and Diff heatmaps (Pearson, Spearman, Kendall, Cramér's V, Phik)
  • Alert changes: new and resolved alerts with trigger values
  • Proportion-based charts for fair comparison across different dataset sizes
  • Interactive HTML report with ECharts (histogram, bar, line, heatmap, scatter, boxplot)
  • Programmatic access: se.diff, .to_html(), .original_report, .synthetic_report
  • Dual ingestion: pl.DataFrame or ProfileReport
  • PEP 561 py.typed marker

Full Changelog: https://github.com/dataxid/dataxid-syntheval/releases/tag/v0.1.0

Assets 2
Loading