Add toggable automatic outlier filtering option using IQR method

[tampueroc]

Problem

Currently, STDF-Viewer allows viewing all numeric data points (e.g., measurements, DUTs) in plots and histograms without an easy way to automatically filter statistical outliers. This can make it difficult to focus on the main data distribution when a few extreme values distort the scale or visual trends.

Expected Behavior

Introduce an optional “Outlier Filter” toggle that automatically hides data points identified as outliers using the Interquartile Range (IQR) method.

The feature should:

• Use a configurable sensitivity parameter (α) that determines how far from the median a point must be to be considered an outlier.
• Apply to numeric visualizations such as histograms, trend plots, and other evaluation views.
• Update visualizations dynamically when the filter setting is changed.

Additional Context

Related to #143

[noonchen]

You are right for histogram, outlier data might mess with bin calculation, and hiding the details of majorities.  I have already added this to roadmap, using same method to detect outliers as boxplot. As for trend plot, user can scale the image on GUI freely without losing any info, in short words, outlier filter will not be applied to trend plots.

…

---Original--- From: "Tom ***@***.***> Date: Thu, Oct 23, 2025 02:38 AM To: ***@***.***>; Cc: ***@***.***>; Subject: [noonchen/STDF-Viewer] Add automatic outlier filtering option usingIQR method (Issue #170) tampueroc created an issue (noonchen/STDF-Viewer#170) Problem Currently, STDF-Viewer allows viewing all numeric data points (e.g., measurements, DUTs) in plots and histograms without an easy way to automatically filter statistical outliers. This can make it difficult to focus on the main data distribution when a few extreme values distort the scale or visual trends. Expected Behavior Introduce an optional “Outlier Filter” toggle that automatically hides data points identified as outliers using the Interquartile Range (IQR) method. The feature should: Use a configurable sensitivity parameter (α) that determines how far from the median a point must be to be considered an outlier. Apply to numeric visualizations such as histograms, trend plots, and other evaluation views. Update visualizations dynamically when the filter setting is changed. Additional Context Related to #143 — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>