Skip to contents

bidux

Overview

The bidux package helps Shiny developers create more effective dashboards using the Behavioral Insight Design (BID) Framework. If you’ve ever wondered why users struggle with your carefully crafted dashboards, or why your beautifully visualized data doesn’t drive the decisions you expected, this package is for you.

The core insight: Technical excellence ≠ User success. Even the most sophisticated analysis can fail if users can’t quickly understand and act on it.

The BID framework bridges this gap by integrating behavioral science, UX best practices, and data storytelling techniques into a systematic 5-stage process:

  1. Interpret the User’s Need - Like defining your research question and understanding your data
  2. Notice the Problem - Like identifying data quality issues or analytical bottlenecks
  3. Anticipate User Behavior - Like checking for statistical biases that could skew results
  4. Structure the Dashboard - Like choosing the right visualization or model architecture
  5. Validate & Empower the User - Like testing your model and ensuring stakeholders can act on results

Installation

You can install the released version of bidux from CRAN with:

And the development version from this repo with:

# install.packages("pak")
pak::pak("jrwinget/bidux")

Usage 

library(bidux)

# Document a complete BID process
process <- bid_interpret(
  central_question = "How are our marketing campaigns performing across different channels?",
  data_story = list(
    hook = "Recent campaign performance varies significantly across channels",
    context = "We've invested in 6 different marketing channels over the past quarter",
    tension = "ROI metrics show inconsistent results, with some channels underperforming",
    resolution = "Identify top-performing channels and key performance drivers"
  )
) |>
  bid_notice(
    problem = "Users are overwhelmed by too many filter options and struggle to find relevant insights",
    evidence = "User testing shows 65% of first-time users fail to complete their intended task within 2 minutes"
  ) |>
  bid_anticipate(
    bias_mitigations = list(
      anchoring = "Include previous period performance as reference points",
      framing = "Provide toggle between ROI improvement vs. ROI gap views"
    )
  ) |>
  bid_structure() |>
  bid_validate(
    summary_panel = "Executive summary highlighting top and bottom performers, key trends, and recommended actions",
    next_steps = c(
      "Review performance of bottom 2 channels",
      "Increase budget for top-performing channel",
      "Schedule team meeting to discuss optimization strategy"
    )
  )

# View implementation suggestions for specific packages
bid_suggest_components(process, package = "bslib")
bid_suggest_components(process, package = "reactable")

# Generate comprehensive reports
bid_report(process, format = "html")
bid_report(process, format = "markdown")

Data-Driven UX with Telemetry

New in 0.3.1: Enhanced telemetry workflow transforms real user behavior data into actionable BID insights.

# Modern approach: analyze telemetry data
issues <- bid_telemetry("telemetry.sqlite")
print(issues)  # Shows organized issue summary with severity levels

# Focus on critical issues using tidy workflows
critical_issues <- issues |>
  filter(severity == "critical") |>
  arrange(desc(user_impact))

# Convert high-priority issues to BID Notice stages
notices <- bid_notices(
  issues = critical_issues,
  previous_stage = interpret_result
)

# Use telemetry flags to inform structure decisions
flags <- bid_flags(issues)
structure_result <- bid_structure(
  previous_stage = anticipate_result,
  telemetry_flags = flags  # Influences layout selection
)

The bid_telemetry() function automatically identifies five key friction indicators:

  • Unused Inputs: UI controls rarely interacted with
  • Delayed Interactions: Long time-to-first-action patterns
  • Frequent Errors: Recurring error patterns that disrupt workflows
  • Navigation Drop-offs: Pages or tabs users rarely visit
  • Confusion Patterns: Rapid repeated changes indicating user uncertainty

Legacy bid_ingest_telemetry() function maintains full backward compatibility while providing enhanced functionality through hybrid objects.

Key Features

Behavioral Science Integration

  • Bias Mitigation: Address anchoring, framing, confirmation bias, and choice overload
  • Cognitive Load Management: Implement progressive disclosure and visual hierarchy
  • Peak-End Rule: Structure experiences for positive, actionable endings

Implementation Support

  • Component Suggestions: Get tailored recommendations for bslib, reactable, echarts4r, and more
  • Layout Selection: Automatic layout recommendations based on content and telemetry flags
  • Concept Dictionary: Access 50+ behavioral science concepts with implementation tips

Real User Data Integration

  • Telemetry Analysis: Transform usage patterns into design insights
  • Evidence-Based Design: Ground design decisions in actual user behavior
  • Iterative Improvement: Track UX improvements with before/after telemetry comparison

Learning More

The BID framework is based on established behavioral science and UX design principles. Explore the concept dictionary:

# Browse all available concepts
bid_concepts() |>
  select(concept, category, description)

# Get detailed information about specific concepts
bid_concept("Processing Fluency")
bid_concept("Hick's Law")

# Search for concepts by keyword
bid_concepts("cognitive") |>
  select(concept, implementation_tips)

Getting Help

Code of Conduct

Please note that the bidux project is released with a Contributor Code of Conduct. By contributing to this project, you agree to abide by its terms.