In 2025, visitors to the High Museum of Art in Atlanta walked into a darkened gallery and found themselves standing in front of three monumental screens. Dense, pulsing images cascaded across them: molecular structures, astronomical data, human genome sequences, climate measurements, all transformed into luminous visual forms synchronized to an electronic score. The work was Ryoji Ikeda's data-verse 1/2/3 (2019-2020), and it used open-source datasets at the thresholds of human knowledge as its raw material. The screens showed not what the data meant but what it looked like when you turned it into images at operatic scale. Ikeda's first museum exhibition in the United States confirmed his reputation as the artist who has done more than anyone to make data visible as an aesthetic phenomenon. Read the review at The Brooklyn Rail.
Data art, also called data visualization art or information art, is art that uses real datasets as its primary medium. The artist takes data from scientific, social, personal, or institutional sources and transforms it into visual, auditory, or physical form. The goal is not to communicate information efficiently, as in scientific data visualization, but to produce an aesthetic and conceptual experience. Data art sits at the intersection of art, science, and design, and it has grown alongside the expansion of big data, machine learning, and computational culture.
This entry covers the history of data art from its origins in 1970s conceptual art to contemporary AI-driven practice, the artists who defined the field, and how data became a material that artists work with the way painters work with pigment.
What Is Data Art and How Does It Work?
Data art is art practice grounded in real data. The artist selects a dataset, processes it through algorithms or visualizations, and presents the result as an artwork. The data can be anything: climate records, stock market transactions, social media posts, genetic sequences, astronomical measurements, migration statistics, or personal logs. What makes the work data art rather than scientific visualization is the artist's intent. A scientific visualization aims to make data legible. Data art aims to make data felt.
A 2025 study published on arXiv analyzed 220 data artworks to understand their design paradigms and intents. The researchers constructed a design taxonomy characterizing techniques including sensation, interaction, narrative, and physicality. They also interviewed twelve data artists about their practice. The study found that artistic data visualization is deeply rooted in art discourse, with distinctive characteristics in both inner pursuits and outer presentations. The research positions data art as a significant domain that has emerged at the intersection of science and art, propelled by the advent of big data. Read the study at arXiv.
The distinction between data art and pragmatic data visualization matters. Robert Kosara, a visualization researcher, drew a spectrum placing artistic visualization and pragmatic visualization at opposite ends. Pragmatic visualization prioritizes clarity and accuracy. Artistic visualization prioritizes aesthetics, emotion, and concept. Data art does not need to be legible. It needs to be meaningful, and meaning in art is not the same as meaning in a chart.
Origins and History
The conceptual roots of data art go back to the 1970s. Kynaston McShine's 1970 exhibition Information at the Museum of Modern Art is widely cited as an early milestone. The exhibition presented works that dealt with information as subject matter, including works that used data, systems, and communication networks. The artists in Information were responding to the information age before that term was common, and the exhibition established the idea that information itself could be artistic material.
In 2007, Fernanda Viégas and Martin Wattenberg explicitly proposed the concept of artistic data visualization. Their Wind Map (2012), which visualized real-time wind patterns across the United States as flowing white lines on a black background, became the first web-based artwork to enter MoMA's permanent collection. Viégas and Wattenberg later led Google's Data Arts Team, founded in 2008 to explore what creativity and technology could do together. Their work established that data visualization could function as art when the visualization prioritized aesthetic and experiential qualities over analytical utility.
The academic community formalized the field. Since 2013, the IEEE VIS conference has included an Arts Program (IEEE VISAP), showcasing artistic data visualizations through accepted papers and curated exhibitions. The program created a venue where data artists could present work to both scientific and artistic audiences, bridging communities that had previously operated separately.
The 2010s and 2020s brought machine learning into data art. Artists began using neural networks not just to visualize data but to process it in ways that produced new visual forms. Refik Anadol's work, which feeds museum collections and urban datasets through machine learning models to produce generative projections, represents this shift. The data is no longer simply visualized. It is processed by an AI system that learns patterns and generates outputs that no human could predict.
Key Artists and Practitioners
Ryoji Ikeda (b. 1966)
Ikeda is the artist most associated with data art at gallery scale. His data-verse 1/2/3 (2019-2020) transforms open-source datasets from NASA, CERN, and the Human Genome Project into monumental audiovisual installations. Earlier works like the planck constant (2018) and point of no return (2018) explore the phenomenological effects of data at the thresholds of perception. Ikeda's practice treats data as raw material for sensory experience, not as information to be decoded. His 2025 exhibition at the High Museum, reviewed in The Brooklyn Rail, confirmed his position as the leading figure in data art.
Refik Anadol (b. 1985)
Anadol uses machine learning to process large datasets into generative visual works. Unsupervised (2022, MoMA) feeds 138,151 works from the museum's collection through a generative model, producing a continuously evolving projection on a 24-foot LED wall. His Machine Hallucinations series processes datasets of astronomical imagery, urban photographs, and architectural records. Anadol is opening Dataland in 2025, described as the world's first museum of AI arts, in the Grand LA complex designed by Frank Gehry. Read more at EVA London 2025.
Fernanda Viégas and Martin Wattenberg
Viégas and Wattenberg created Wind Map (2012), the first web-based artwork in MoMA's permanent collection. The work visualized real-time wind patterns as flowing lines, creating an image that was both scientifically accurate and aesthetically mesmerizing. They led Google's Data Arts Team and created The Shape of Song (2001), which visualized the structure of musical compositions as arching lines connecting repeated passages.
Anna Ridler (b. 1986)
Ridler's Every Single Iris on the Internet (the first 100,000 images) (2025) draws on all instances of the word "iris" found within the LAION dataset, revealing how large-scale machine learning systems are built on fragments of an internet that is already disappearing. As the work cycles through images of flowers, eyes, and personal names, broken links and missing files gradually fade until the screen turns black. The work is both data art and a critical examination of the datasets that underpin AI systems. Read more at annaridler.com.
Lisa Autogena and Joshua Portway
Autogena and Portway created Most Blue Skies (2007-2009), a work that used real-time environmental data from cities worldwide to determine which city had the bluest sky at any given moment. The work combined data analysis with poetic intent, turning atmospheric measurements into a meditation on beauty, environmental monitoring, and the quantification of the natural world.
Albert-László Barabási
Barabási, a network scientist and Fellow of the American Physical Society, runs a data art lab at Northeastern University. His work visualizes complex networks, from scientific collaborations to cellular metabolism. Barabási has argued that data has become a vital medium for artists dealing with societal complexity, stating that traditional tools of art are inadequate for the challenge and that artists must embrace new tools that scale to the complexity of contemporary life.
Data Art Today
Data art in 2026 is shaped by two forces. First, the sheer volume and availability of data. Open government datasets, real-time environmental sensors, social media APIs, and scientific databases provide more raw material than any artist could process in a lifetime. Second, the integration of machine learning. AI models can find patterns in datasets that are invisible to human analysis, and data artists use these models to generate visual forms that emerge from the data's internal structure rather than from the artist's predetermined visualization scheme.
The field has also developed a critical dimension. Artists like Anna Ridler examine the datasets themselves, revealing their gaps, biases, and ghosts. The LAION dataset that Ridler explores in Every Single Iris contains material that is frequently a decade old, much of it degraded, with broken links and missing files. This critical data art asks not what data can show us but what data hides, what it forgets, and what it erases through its collection.
See Data Art in Person
Data art is best experienced at scale, where the density and rhythm of the data become physically immersive. Look for Ryoji Ikeda's installations, which have been exhibited at the High Museum, the Centre Pompidou, and the Carriageworks in Sydney. Refik Anadol's Unsupervised is on continuous display at MoMA in New York. For web-based data art, explore the Wind Map archive online and Anna Ridler's portfolio at annaridler.com.
For more on the broader field, read our entries on generative art, digital art, and creative coding, or explore our blog post on how to read a painting to learn more about analyzing visual art across different mediums.
