When the Chart Lies: RDA Radar Chart Scaling Decisions

A nutritionist builds a radar chart to compare 2 households' diet quality across 7 nutrients. Using absolute intake (grams), Household A's chart looks nearly full while Household B's looks sparse. But when she switches to percentage of RDA, the picture reverses — Household A is actually deficient in several micronutrients while Household B, who eats smaller volumes of diverse foods, is more balanced. The analyst must manually trace both scalings to understand why the chart flips.

# Aplly.xyz Case Study Submission

## Title
When the Chart Lies: RDA Radar Chart Scaling Decisions

## Type
Data Analytics

## Difficulty
Advanced

## Estimated Time
75 minutes

## Overview
A nutritionist builds a radar chart to compare 2 households' diet quality across 7 nutrients. Using absolute intake (grams), Household A's chart looks nearly full while Household B's looks sparse. But when she switches to percentage of RDA, the picture reverses — Household A is actually deficient in several micronutrients while Household B, who eats smaller volumes of diverse foods, is more balanced. The analyst must manually trace both scalings to understand why the chart flips.

## Case Details

Function Focus: Visualization scaling sensitivity analysis, radar chart axis design, percentage-of-RDA vs absolute scaling, chart integrity auditing

Scenario:
A nutrition researcher builds a radar chart comparing 2 households' daily nutrient intake across 7 dimensions. Using absolute intake in grams, Household A's chart area looks nearly full — they eat large volumes of rice and dal. Household B's chart area looks tiny — they eat smaller portions. The researcher concludes Household A has better diet quality. But a colleague suggests switching each axis to percentage of RDA instead of absolute grams. When she does, the picture reverses: Household A drops below 100% on 3 micronutrients, while Household B, eating smaller but more diverse portions, meets or exceeds RDA on 6 of 7 nutrients. The analyst must manually compute both chart scalings for both households and determine which scaling produces a truthful, not misleading, visual comparison.

Dataset Structure:
- 2 households with daily intake of 7 nutrients (in grams or mcg)
- RDA values for each nutrient
- Two scaling methods: absolute intake (grams) vs percentage of RDA

Tasks:
1. For each household, convert the absolute intake of all 7 nutrients to percentage of RDA — show the arithmetic for at least 3 nutrients
2. For Household A, compare the radar chart shape under absolute scaling vs % RDA scaling — which nutrients drive the shape in each version? Identify the nutrient that changes the most between scalings and explain why
3. For Household B, do the same comparison — identify which nutrient appears worst under absolute scaling but looks adequate under % RDA scaling, and explain why
4. Based on your analysis, recommend which scaling the researcher should use for a diet quality comparison — justify with specific nutrient examples from the 2 households, and state what the absolute-scaled chart hides that the % RDA chart reveals
5. Only after submitting your manual analysis, re-run the percentage conversions in a spreadsheet or AI tool and report any discrepancies — then propose a 3rd scaling method (e.g., log scale or normalized range) that could complement the % RDA view

Expected Output:
A visualization integrity memo containing: (a) % RDA conversion table for both households with sample calculations, (b) Household A shape analysis comparing both scalings with the most-changed nutrient identified, (c) Household B shape analysis with the most-improved nutrient identified, (d) a scaling recommendation with specific justification, and (e) a round-trip discrepancy note plus a proposed 3rd scaling method.

Evaluation Criteria:
Correct % RDA arithmetic for all 7 nutrients × 2 households, accurate identification of the most-changed and most-improved nutrients under scaling change, a well-justified scaling recommendation based on specific nutrient examples, meaningful proposal for a complementary 3rd scaling, and honest discrepancy reporting.

## Data Sources

Daily Nutrient Intake — Household A (large volumes, rice-dal based):
| Nutrient | RDA (adult female) | Household A Intake (absolute) | Unit |
|---|---|---|---|
| Energy | 2,100 | 2,350 | kcal |
| Protein | 55 | 52 | g |
| Iron | 21 | 14 | mg |
| Vitamin A | 600 | 420 | mcg RAE |
| Calcium | 800 | 720 | mg |
| Vitamin C | 80 | 35 | mg |
| Fiber | 25 | 18 | g |

Daily Nutrient Intake — Household B (small portions, diverse foods):
| Nutrient | RDA (adult female) | Household B Intake (absolute) | Unit |
|---|---|---|---|
| Energy | 2,100 | 1,680 | kcal |
| Protein | 55 | 50 | g |
| Iron | 21 | 19 | mg |
| Vitamin A | 600 | 780 | mcg RAE |
| Calcium | 800 | 640 | mg |
| Vitamin C | 80 | 85 | mg |
| Fiber | 25 | 22 | g |

Scaling Methods:
1. Absolute scaling: each axis uses raw grams/mcg. Range = 0 to max observed across both households per nutrient.
2. % RDA scaling: each axis uses (intake ÷ RDA) × 100%. Range = 0% to 150% (nutrients exceeding RDA are capped at 150% for visual consistency).

Context:
- Researcher's initial chart (absolute scaling) covered 7 axes: Energy, Protein, Iron, Vit A, Calcium, Vit C, Fiber
- Chart footnote: "Axis scales are independent — each nutrient has its own max value"
- Colleague's concern: "Independent axes make the chart shape meaningless for comparison"

## Solution Frameworks
Visualization scaling analysis, radar chart axis design, percentage normalization, chart integrity auditing, comparative visual perception

## Solver Guidance & Tutorials
Link to: "Visualization Integrity — When Chart Scaling Misleads" tutorial

## What You'll Learn
- How axis scaling decisions change the apparent story of a radar chart
- Converting absolute values to percentage of RDA for meaningful comparison
- Detecting when independent axis scales create a misleading visual
- Choosing the right scaling method for diet quality visualization

## Tags
radar chart, data visualization, RDA scaling, chart integrity, nutrition analytics

## Registration Links
- Register as Solver
- Register as Evaluator

Daily Nutrient Intake — Household A (large volumes, rice-dal based):
| Nutrient | RDA (adult female) | Household A Intake (absolute) | Unit |
|---|---|---|---|
| Energy | 2,100 | 2,350 | kcal |
| Protein | 55 | 52 | g |
| Iron | 21 | 14 | mg |
| Vitamin A | 600 | 420 | mcg RAE |
| Calcium | 800 | 720 | mg |
| Vitamin C | 80 | 35 | mg |
| Fiber | 25 | 18 | g |

Daily Nutrient Intake — Household B (small portions, diverse foods):
| Nutrient | RDA (adult female) | Household B Intake (absolute) | Unit |
|---|---|---|---|
| Energy | 2,100 | 1,680 | kcal |
| Protein | 55 | 50 | g |
| Iron | 21 | 19 | mg |
| Vitamin A | 600 | 780 | mcg RAE |
| Calcium | 800 | 640 | mg |
| Vitamin C | 80 | 85 | mg |
| Fiber | 25 | 22 | g |

Scaling Methods:
1. Absolute scaling: each axis uses raw grams/mcg. Range = 0 to max observed across both households per nutrient.
2. % RDA scaling: each axis uses (intake ÷ RDA) × 100%. Range = 0% to 150% (nutrients exceeding RDA are capped at 150% for visual consistency).

Context:
- Researcher's initial chart (absolute scaling) covered 7 axes: Energy, Protein, Iron, Vit A, Calcium, Vit C, Fiber
- Chart footnote: "Axis scales are independent — each nutrient has its own max value"
- Colleague's concern: "Independent axes make the chart shape meaningless for comparison"

Visualization scaling analysis, radar chart axis design, percentage normalization, chart integrity auditing, comparative visual perception