How Scientists Turn 24/7 Movement Data Into Charts That Actually Make Sense

Every step you take, every minute you spend sitting, and every hour you sleep is now potentially recorded by a small accelerometer strapped to your wrist or hip. Wearable sensors have transformed human movement research, generating torrents of continuous data that capture the full 24-hour cycle of physical activity, sedentary behaviour and sleep. Yet according to a new study from the European LABDA project, the field has a glaring blind spot: almost nobody has systematically studied how all this data should be visualised and communicated. The research, published in the Journal of Activity, Sedentary and Sleep Behaviors, offers the first comprehensive map of the metrics scientists use to summarise movement behaviour and the charts they draw to explain them.

The team, led by Marian Marchiori of the University of Southern Denmark, conducted an umbrella review, a systematic search of reviews, to catalogue the ways accelerometer data are condensed into meaningful numbers. Searching Scopus and Web of Science in February 2025, they screened 324 reviews and ultimately included 93, published between 2012 and 2024. Those reviews collectively covered 5,667 original studies, of which 4,669 relied on accelerometers. From this enormous body of work, the researchers identified 134 distinct output metrics used to describe human movement behaviour, a striking illustration of how fragmented measurement practice has become in a field that ostensibly studies the same handful of behaviours.

The most familiar metrics dominated the landscape. Step counts appeared in 67 of the 93 reviews, making it the single most cited measure, followed closely by time spent in moderate-to-vigorous physical activity, or MVPA, which appeared in 65 reviews. Time in light physical activity, raw accelerometer counts, total sedentary time and total physical activity time rounded out the top tier. But beyond these household names, the long tail of metrics reveals genuine complexity: measures of bout duration, fragmentation, circadian patterns, intensity gradients and machine-learned activity types all jostle for space in the literature, each capturing a different facet of how people move through their days.

To bring order to this sprawl, the researchers performed an adapted thematic analysis, grouping the 134 metrics into five conceptual themes: relative measures, frequency, absolute time, point in time, and summary, pattern or derived measures. These categories map neatly onto the questions researchers actually ask of their data. Frequency metrics answer the question of how often a behaviour occurs, such as how many times a day someone stands up from sitting. Absolute time metrics answer how long a behaviour lasts. Relative measures capture what proportion of the day is spent in each behaviour, while point-in-time metrics pinpoint when behaviours begin and end. Summary and derived measures, such as raw acceleration aggregates like ENMO or MIMS units, describe the overall quantity and intensity of movement.

The visualisation side of the story proved even more revealing. When the researchers attempted a systematic search for studies on visualising accelerometer metrics, they came up essentially empty-handed; the literature simply does not treat visualisation as a topic worthy of systematic investigation in its own right. Falling back on non-systematic web searches and consultations with the generative AI tool Microsoft Copilot, they identified 28 different types of visualisations applicable to the metrics they had catalogued. Yet in practice, researchers overwhelmingly rely on just three: bar charts, line graphs and pie charts, typically chosen without any explicit rationale for why one format suits a particular metric or message better than another.

This gap matters more than it might seem. Data visualisation is not a cosmetic afterthought; it shapes how findings are understood, and the optimal chart for one audience may fail another entirely. A policy maker glancing at a report, a clinician interpreting a patient’s activity profile and a researcher scrutinising statistical patterns bring different expectations and different levels of statistical literacy. Terms like MVPA, second nature to exercise scientists, mean little to the general public. The authors point out that presenting only point estimates without any indication of variability can actively mislead audiences without statistical training, and that the same data rendered in different visual forms can elicit competing interpretations.

To address the problem, the team built a conceptual framework grounded in the sender-receiver model of communication, borrowed from the philosophy of language. In this framing, the researcher is the sender, the target audience is the receiver, and the visualisation is the coding strategy that transforms raw data into a transmissible message. The data themselves constitute the message, while journals and conference presentations serve as the channel. Noise, in this model, is anything that obscures the intended meaning, from cluttered graphics to jargon-heavy labels. The framework walks researchers through a sequence of decision points: identify whether the underlying data are unclassified, producing intensity or quantity metrics, or classified, producing behaviour types and intensity levels; determine which research question the metric answers, whether how much, how long, how often, when, or what proportion; and only then select a visualisation suited to that category and context.

The framework is deliberately non-prescriptive. Rather than dictating that a given metric must always be drawn a certain way, it organises options and illustrates pathways. One worked example contrasts a pie chart showing daily proportions of movement behaviours with a stacked bar plot, which preserves information about day-to-day variation and temporal patterns while still allowing comparison between averages and individual days. Another contrasts a simple bar plot counting activity bouts with an event plot, which additionally reveals the duration and timing of each individual bout. In both cases, the less conventional option can convey substantially more insight, depending on what the researcher actually wants to communicate.

The study’s authors are candid about its limitations. No review protocol was pre-registered, the search captured reviews published only up to 2024, and the exploratory visualisation searches, by necessity, were not systematic. The framework also lacks empirical validation: nobody yet knows how different audiences actually perceive the visualisations it recommends. The team calls for future work involving co-design with end users, comprehension testing and iterative refinement, ideally extending the framework to offer audience-specific guidance. They also note that the field would benefit enormously from a common taxonomy of movement behaviour metrics and from harmonised accelerometer data processing, initiatives already underway in consortia such as ProPASS, which would make visualisations more comparable across studies.

Nevertheless, the significance of the work lies in reframing visualisation as a communication problem rather than a technical one. With the World Health Organization reporting that insufficient physical activity has been rising globally since 2000, with stark disparities by sex, age and geography, the stakes of clear communication are high. Physical activity guidelines can only change behaviour if their messages land, and charts are often the first and most persuasive point of contact between research and the public. By cataloguing 134 metrics, mapping 28 visualisation types onto them, and embedding the whole exercise in a theory of communication, the LABDA team has given movement scientists something the field has never had: a systematic starting point for deciding not just what to measure, but how to show it.

Subject of Research: Visualisation of accelerometer-based 24/7 human movement behaviour metrics

Article Title: Visualising accelerometer-based 24/7 human movement behaviour data: an umbrella review and framework development from the LABDA project

Article References: Visualising accelerometer-based 24/7 human movement behaviour data: an umbrella review and framework development from the LABDA project. (n.d.). https://doi.org/10.1186/s44167-025-00088-6

Image Credits: AI Generated

DOI: 10.1186/s44167-025-00088-6

Keywords: accelerometers, physical activity, sedentary behaviour, sleep, data visualisation, wearables, 24/7 movement behaviours, MVPA, step counts, umbrella review, LABDA project, science communication