The Project
This project examines whether higher prices correspond to more complex formulations or whether consumers are primarily paying for brand prestige. Every product is scored using the same two-part methodology: an Ingredient Quality Score that evaluates the formula itself, and a Price Fairness Multiplier that asks whether the cost is justified.
Using a custom-built dataset and generative visualization techniques, skincare formulations will be translated into organic, petri-dish-inspired digital ecosystems where ingredients function as particles within compositional systems. The final outcome aims to merge data analysis with immersive editorial-style visualization.
Ingredient Quality Score — IQS
The IQS captures what is actually in the formula. It rewards beneficial ingredients, penalises harmful ones, and normalises for list length so longer formulas are not automatically rewarded.
Base Formula
Icky Ingredient Position Weighting
Cosmetic ingredient lists are legally required to be ordered by descending concentration. An icky ingredient near the top of the list is far more concerning than one buried near the bottom.
Base Position Weight
Position 5 of 40 → 0.875 (heavy). Position 38 of 40 → 0.05 (near negligible).
The 1% Rule & Floor of 0.25
Regulations allow ingredients at ≤1% to be listed in any order. Without a floor, icky ingredients listed late would receive near-zero penalties — even though their presence at any concentration still carries risk for sensitive skin.
0.25 ensures any icky ingredient carries at least a quarter of the concern of a prominently listed one — high enough to matter, low enough to still reward products that bury them late.
Final Weighted Sum
Price Fairness Multiplier — PFM
The PFM adjusts the final score based on price per ml. A great formula at a budget price is rewarded. A poor formula at a luxury price is penalised. Accessible-tier products are unaffected.
PFM Formula
IQSnorm rescales each product to a −1 to +1 range across the full dataset, so every product is judged relative to the others.
Ingredient Weights
| Rating | Weight | Rationale |
|---|---|---|
| Superstar | × 3 | Rare, high-impact actives — strongest positive signal in a formula. |
| Goodie | × 1 | Beneficial ingredients — positive but baseline contribution. |
| Icky | × 2 | Harmful or questionable — penalised harder than goodies are rewarded, scaled by position. |
| Neutral | × 0 | Functional base ingredients — no meaningful positive or negative signal. |
Position Weighting — Example (40 ingredients)
| Position | Raw weight | With floor | Notes |
|---|---|---|---|
| 3 | 0.925 | 0.925 | Heavy penalty |
| 10 | 0.750 | 0.750 | Significant |
| 20 | 0.500 | 0.500 | Moderate |
| 30 | 0.250 | 0.250 | Floor reached |
| 35 | 0.125 | 0.250 | Floor kicks in |
| 38 | 0.050 | 0.250 | Floor kicks in |
| 39 | 0.025 | 0.250 | Floor kicks in |
Price Tiers
| Price / ml | Tier | Sensitivity | Label |
|---|---|---|---|
| ≤ $0.20 | Budget | +0.15 | Very Affordable |
| $0.21 – $0.60 | Mid-range | 0.00 | Accessible |
| $0.61 – $1.50 | Premium | −0.20 | Elevated |
| > $1.50 | Luxury | −0.35 | High-End |
PFM Scenarios
| IQS | Price | Result |
|---|---|---|
| High | Budget | PFM > 1 — rewarded |
| Low | Luxury | PFM < 1 — penalised |
| High | Luxury | PFM ≈ 0.72 — neutral |
| Low | Budget | PFM ≈ 0.88 — forgiving |
Final Score
Final scores are normalised to 0 – 100 across the dataset. Grade bands reflect relative performance, not an absolute standard.
Grade Bands
Ingredient Categories
Superstar
Rare, high-impact actives. The strongest positive signal in a formula. Weighted ×3.
Goodie
Beneficial ingredients with a positive but baseline contribution. Weighted ×1.
Icky
Harmful or questionable. Penalised harder than goodies are rewarded, scaled by position. Weighted ×2.
Neutral
Functional base ingredients with no meaningful positive or negative signal. No score impact.
Data Source
Ingredient lists and ratings sourced from Inci-Decoder. Classifications mapped to a custom superstar / goodie / icky / neutral system. All scoring calculated manually.