Before the formula. Know the material.

A useful search starts with a development decision.

The working public-data subset below lets you select 18 preparation–experiment cases from two studies, move between the preparation, method and observations, compare reported fields and download an evidence passport. Fifteen cases retain the 165 trial medians from the same 11 participants across different liquids. Its purpose is to sharpen a development question while keeping the measured conditions and source limits attached. The wider collection and commercial service remain proposed.

The proposed atlas begins with a question: should a material enter a film study, a water-handling comparison, a compatibility screen or a sensory investigation? The viewer selects the physical system, the intended endpoint and the condition that matters. Those choices narrow the evidence to experiments that could inform that decision. A supplier name or ingredient family remains useful, but neither substitutes for the identity of the preparation actually tested.

Each selected record would open into a full evidence workspace. The workspace keeps the original result beside its method, then shows the missing bridge to the reader’s proposed use. Its practical output is a comparison brief: what can already be compared, what must be measured again, and what would change the decision. The initial edition would focus on a small number of thoroughly documented material systems rather than an apparently exhaustive catalogue.

Compare the basis of the measurement before comparing its magnitude.

The proposed comparison view would put two records side by side and mark each field as matched, different, unreported or inapplicable. The fields would include material state, concentration basis, medium, temperature, time, instrument and endpoint definition. No overall quality score would collapse these differences into a reassuring number. A matched unit does not make two experiments equivalent, and an unreported condition cannot be treated as a match.

For a water-handling question, a deposited polymer film and an isolated EPS preparation may provide complementary reasoning without permitting a numerical ranking. Borrmann et al. (2022) study PVP-based films with attention to sorption and relaxation; their coefficients would remain attached to that system. The reader could use the comparison to design a common film assay, while the atlas would preserve the distinction between a plausible experimental bridge and an observed result.

(Borrmann et al, 2022)

Use the existing friction analysis as a connected evidence case.

For a spreading or touch question, the atlas would link the material record to the existing tribology analysis and retain its participant, liquid, speed and load information. Fehlberg et al. (2025) provide repeated observations rather than a generic ingredient rating. Readers would see which observations supported the published model and which remain outside that fit selection. The interface would make the role of geometry and rheological assumptions visible before offering any derived comparison.

The decision brief would distinguish measurable friction from a claimed sensory outcome. A formulator could ask whether a candidate changes spreading resistance under a defined protocol, then specify a separate sensory study if perception is the commercial endpoint. The atlas would not turn friction into absorption, penetration or moisturisation. Its value is to make the next question testable and to prevent a convenient physical metric from silently becoming a consumer claim.

(Fehlberg et al, 2025)

Every transformation needs a visible origin.

Each record would retain its author–year citation, original URL, source location, units, preparation identity and reuse status. Downloadable numerical records would separate observations, published summaries and Bio’s reproduced calculations. Corrections would create a new version and a change note, preserving the earlier record for reproducibility. The public interface would show a checked date without implying that a historical dataset updates itself.

This is a proposed data experience, not a compatibility certification or an ingredient database with finished-formula approval. The first build would reuse carefully checked public cases and introduce further records only after source, licence and numerical review. Readers with a specific material or formulation decision could discuss the question and what evidence would be needed. That discussion would establish the scope before promising testing, analysis delivery or commercial performance.