Documentation Validation Request: Confidence Intervals in Sensor Data Visualization

[mosoriob]

Overview

This is a request to validate the technical accuracy and clarity of documentation related to confidence intervals as displayed in the LineConfidenceViz component. The documentation explains statistical concepts that are critical for users to properly interpret visualization output.

https://dso-tacc.netlify.app/docs/confidence-explanation

Documentation to Review

The content covers:

• Definition of confidence intervals (95% confidence level)
• Two types of intervals implemented:
  • Percentile-based intervals (non-parametric)
  • Parametric intervals (t-distribution based)
• T-value selection based on sample size

Validation Requirements

Technical Accuracy

• Verify that the formula description for parametric intervals is correct: mean ± (t-value × standard error)
• Confirm t-values are appropriate for the specified sample size ranges:
  • 30+ data points: 1.96
  • 20-29 data points: 2.09
  • 10-19 data points: 2.23
  • Fewer than 10 points: 2.58
• Validate that percentile bounds (2.5th and 97.5th) correctly represent a 95% confidence interval

Implementation Alignment

• Confirm the LineConfidenceViz component actually implements both methods as described
• Verify the data structure includes all mentioned fields: lower_bound, upper_bound, parametric_lower_bound, and parametric_upper_bound
• Check that the visualization correctly displays these intervals as shaded areas

Documentation Clarity

• Evaluate if the explanation is understandable for our target audience
• Identify any technical terms that need further explanation
• Suggest improvements for clarity or completeness

Reviewers

• @wmobley

[mosoriob]

We need to validate the confidence interval calculations in our sensor viz docs. Pawel's team has the right expertise for this - can they verify the t-values, component implementation, and clarity of explanation?