textile × data × perception

Where
threads
become
data

An exploration of tactile information — the intersection of woven structure and humanistic data visualization, in the tradition of Giorgia Lupi's belief that data has a human face.

↳ Each thread is a measurement.
Each knot, a data point.
Warp and weft as x and y.
Tension as standard deviation.
§ 02 — Method

How information becomes material, and material becomes meaning

01 / COLLECT

Gathering the weft

Data is collected from ambient, ecological, and personal sources — sleep rhythms, urban noise levels, microbiome diversity. Each dataset becomes a palette of thread colors and tensions.

02 / ENCODE

Mapping to material

Variables are assigned to physical properties: color saturation maps to data magnitude, thread weight to frequency, crossing density to correlation strength between two observed phenomena.

03 / WEAVE

The loom as algorithm

The act of weaving executes the encoding. No two passes of the shuttle are identical — human variation becomes an aesthetic feature, a departure from machine precision that renders data humanistic.

I
§ 03 — Philosophy

Data is not cold. It is gathered by hands, about lives.

In the tradition of Giorgia Lupi's "data humanism," we believe that the visual representation of information should retain the fingerprint of its human origin. Statistics are not abstract — they are counts of people, measurements of growth, recordings of the world as someone experienced it on a specific day.

Textile art embodies this philosophy physically. A woven piece remembers every hand that touched the loom, every decision made about color and tension. It is inherently imprecise, contextual, personal — and that is precisely what makes it true.

Our work sits at the intersection of these two traditions: rigorous data collection married to tactile, imperfect, beautiful material encoding.

All datasets are treated as continuous observations rather than discrete values. Interpolation is applied before encoding to produce smooth gradient transitions in the woven surface.

Human loom operation introduces ±3–7% variation in thread tension. This is considered signal, not noise. It is the mark of authorship.

Merino wool / 120 tex
Raw silk / 40 denier
Linen warp / 16/1 lea
Cotton supplementary weft

§ 04 — Technique

The four languages of data weaving

Technique I

Ikat encoding

Threads are resist-dyed before weaving according to a data distribution. The resulting bleed patterns mirror the natural uncertainty in ecological measurements.

Technique II

Supplementary weft

Additional threads float above the base cloth to annotate outliers — the Lupi-esque act of circling and naming the exceptional datum.

Technique III

Pile and texture

Raised pile encodes intensity — areas of high measurement density emerge physically from the surface, legible by touch as much as by sight.

Technique IV

Color sequence

A handcrafted palette of 47 dye baths, each corresponding to a quantile in the source dataset, applied sequentially to yarn before winding.