Deep within the underbellies of online forums and social media sites, a global network of passionate knitters, crocheters, sewists and stitchers are quietly pioneering a bespoke form of data visualization, conveying complex statistical ideas through artistic handmade crafts.
These “data textiles” 一 so called by researcher Sydney Purdue of Northeastern University’s Khoury Vis Lab, who surveyed this secret community with her collaborator Eduardo Puerta 一 come in all shapes and sizes, and can deal with topics ranging from climate change and gun violence to the crafter’s personal success rate on their daily Wordle games.
Perhaps the simplest form of data textile is a temperature blanket: a weather history visualization in which different colors of yarn connote hotter or colder temperatures in a given area. Each hand-sewn stripe on a temperature blanket represents a single day’s heat or coolness. After a full calendar year, the result is a cozy, rainbow-patterned work of art that is also as scientifically sound as a traditional line graph, Purdue said.
More than a clever way to mix hobby with statistical rigor, data textiles have the potential to unlock entirely new disciplines within the data visualization field, she said. Further study of the art form could allow Northeastern researchers to explore fresh ideas about how to represent data in a way that better taps into people’s personal values and emotions, she said.
Hand-stitched items that a person can wear, snuggle with, give to someone as a gift or even decide to unravel may have the power to stir up emotional connections to data that even the most accurate chart or diagram could never replicate
“There’s something different about physical visualizations. You can touch them, you can hold them, you can live with them, you can build them slowly,” said Melanie Tory, a Northeastern professor of practice with appointments in the Khoury College of Computer Sciences (Khoury) and the College of Arts, Media and Design (CAMD). “And when you get into textiles, that gets super interesting, because those are artifacts that live in your world.”
But when Purdue and Puerta first raised the idea to explore the largely uncharted world of data textiles in 2024, Tory said she and fellow jointly-appointed professor Enrico Bertini had some doubts about the topic’s relevance to their field. The concept was far from the mainstream visualization work performed at the Khoury Vis Lab, she said, which typically focuses on the more traditional tables, numbers and graphs that were once referred to as information visualization.
Purdue and Puerta sought out data textile creators using snowball sampling, a non-probability recruitment technique that involved contacting local Boston yarn shops, craft stores, online forum members and other organizations, who in turn circulated the research idea to their customers and colleagues. They soon came back to Tory and Bertini with over 200 responses.
“We were honestly floored,” Tory said. “[Bertini] and I had no idea that this was such a rich space of creators who were really passionate about their work and wanted to show it, and hadn’t engaged with the academic visualization community really at all.”
Purdue said she combined her skillset from her interdisciplinary design and media program in CAMD with Puerta’s computer science background in Khoury and got to work cataloguing the submissions into a novel data textile archive.
Some participants created quilts, jackets or scarves, Purdue said, while others used embroidery, lacework or beading. One cross stitch project, made by Glenda Adams simply for fun, used colorful squares to track her and her wife’s Wordle guesses over a month’s time. Juliette Pécaut, meanwhile, knitted a scarf with thin stripes representing the daily severity of her migraine symptoms. And a trio of crocheted tapestries by Madison Snell represented the three chapters of her master’s thesis, with different colors and stitch types on each row indicating whether a thesis sentence had citations, keywords or reference figures.
“Some of these were memorial projects that helped people process grief. And some were very clearly intended for display, or for activism, or for science communication,” Purdue said.
There were crafters who told Purdue and Puerta that working slowly and deliberately on the projects helped them manage their anxiety or ADHD symptoms, and others who said they hated the process because of its repetitive, time-consuming nature, and vowed never to make one again.
Tory said the breadth of craft types and data topics sparked a number of intriguing questions about the potential benefits of collaboration between the Khoury Vis Lab and data textile creators. Those creators may learn things that change the way they practice their art, and in turn, Northeastern researchers can explore their methods of designing data-based media that resonate with laypeople, she said. The majority of the data visualization field considers subject matter experts to be the target audience when creating graphics, Tory said, but rarely do they think about how best to embed data in the lives of regular people.
Purdue and Puerta’s research also highlighted the merit of “slow visualization,” Tory said. A typical problem tackled by the data visualization field is making data patterns understandable to people as quickly and seamlessly as possible, like a pixel-perfect bar graph with clear labels, she said.
“But by knitting a blanket or crocheting or quilting, you’re engaging with your data in a super slow manner, and there is something valuable in terms of personal meaning, in actually spending the time with the data,” Tory said.
The research has inspired Tory to pursue additional projects that incorporate data textiles and slow creation practices into other communities. For example, in the fall semester, Tory will host data crafting workshops for people with Type 1 diabetes, many of whom wear glucose monitors and other data-tracking devices to manage their illness.
“One of the things that we’re curious about is whether using some of these physical media to explore their data might be useful to them in some way,” Tory said. “These are people who are not data textile creators, but maybe we can use that as a tangible way of looking at their data.”
