Visualizing complex datasets in Edge Impulse

Edge Impulse is a developers-first company enabling hobbyists, data-scientists, and hardcore engineers to build sophisticated machine learning models to detect rich events across sensors, MCUs, and other IoT devices. With the ease of use, data richness, and by capturing 90% of data previously discarded, we help people build better, and more useful products that resonate with humans and machines alike. In this blog post, we'll show some really great recent improvements to our data visualizations, including how to detect and update mislabeled data, validate your signal processing pipeline parameters, and how you compare test data with your training set.

This is a companion discussion topic for the original entry at