Developer and Research Data Access
Continuous, real-world plantar pressure and gait data is one of the scarcest resources in movement science. Pedisteps offers researchers and developers access to structured, longitudinal sensor data under a data-use agreement, not a public API, instead of another small, one-time lab dataset.
The Data Bottleneck in Gait Research
If you’ve worked on gait analysis, fall-risk modeling, or plantar-pressure machine learning, you already know the uncomfortable truth: most publicly available datasets are small, collected in a single lab session, on a narrow population, using specialized equipment nobody wears outside a research building. Reproducing results across datasets is hard because there simply are not many datasets, and the ones that exist rarely reflect real-world, longitudinal, everyday walking.
The field has started to respond, newer benchmark datasets have pushed toward wider participant pools than earlier plantar-pressure collections, but even these remain point-in-time, in-lab captures rather than continuous, real-world signal collected over weeks or months in a subject’s actual daily life. That gap between “one clean lab session” and “how someone actually walks over time” is exactly what limits progress in fall-risk prediction, rehabilitation outcome modeling, and gait-based digital biomarkers.
Pedisteps was built as a continuous data-collection platform from day one, not a lab-only device. That means the underlying dataset is ongoing plantar-pressure and gait data from real people walking in real shoes in the real world. Access for research or development is by inquiry and a signed data-use agreement. There is no public API.
What's available for research
1. Structured sensor streams. Plantar pressure distribution, step-level timing, cadence, gait symmetry, weight-bearing ratios and balance-relevant metrics, available as structured, timestamped export rather than raw unprocessed waveforms you have to reverse-engineer.
2. Longitudinal, not one-shot. Because Pedisol insoles are worn continuously in daily life rather than for a single lab session, the resulting data supports longitudinal analysis: trend detection, intervention response modeling, and within-subject change over time.
3. Consent-based access. Data access for research and development use is structured around explicit participant consent and a data-use agreement. There is no public API, public Unity SDK, or published schema.
4. Shared under agreement. Sensor fields, sampling rates and derived metrics are shared with qualifying projects under a data-use agreement, not published as a public schema or self-serve API.
Request research access
Why Build on Pedisteps Data
Real-world signal, not lab artifacts. Gait captured during a 10-meter clinic walk test behaves differently than gait captured across a normal day of errands, stairs, uneven ground and fatigue. Models trained only on lab data often underperform in deployment for exactly this reason.
Scale that small studies can’t match. Instead of the handful of subjects and single sessions typical of legacy plantar-pressure datasets, continuous collection across a growing user base means larger, more diverse, and more statistically powered datasets over time.
Built for fall-risk and rehab outcome modeling. The metrics Pedisteps captures, symmetry, weight-bearing distribution, cadence variability, balance-relevant pressure shifts, map directly onto the features most used in published fall-risk and rehabilitation-outcome models, so integration work is minimal.
A path from research to product. Because the same sensor platform already powers a live clinical and consumer product, validated models built on Pedisteps data have a direct path to deployment, not just a paper.
The State of Plantar-Pressure and Gait Datasets
Historically, plantar-pressure and gait datasets used for machine learning research have been small and narrow, often a few dozen participants captured in a single in-lab session with a pressure mat or an instrumented walkway. That has made it hard to build models that generalize, and hard to benchmark new approaches against a meaningfully sized, diverse population.
More recent efforts in the field have introduced larger benchmark datasets aimed at addressing exactly this scarcity, expanding participant counts well beyond earlier collections. That’s a meaningful step forward, but these remain snapshot datasets: one visit, one walk, one moment in time per participant.
Continuous, wearable-based collection is the next step the field has been pushing toward, longitudinal data that captures how gait and pressure patterns evolve with fatigue, time of day, intervention, and recovery, not just how they look in a single controlled trial. That’s the gap Pedisteps’ data platform is built to fill for researchers and developers who want to go beyond another one-time lab dataset.
Research and developer access is by inquiry. There is no public API or public Unity SDK. Qualifying projects can receive data export under a signed data-use agreement.
Pedisteps research access
Inquiry and data-use agreement only
Resources: read more about the VRsteps platform and research background, or contact us to discuss a data-use agreement for your project.
Frequently Asked Questions
Ready to see Pedisteps in action?
Stay Connected with Pedisteps
The Pedisteps mobile app helps researchers and study teams keep insole setup simple, capture everyday walks and support consistent data collection.
Use the GaitIQ dashboard to review gait speed, symmetry and balance trends, then bring structured context into your research workflow.
