All Projects

Visualizing Cognitive Data

How I designed complex data visualizations for a type of data no one had seen before

Product Designer

12 weeks

Web Design

The Story

Pison had launched an app with baseball teams to test how our watch's motion sensitivity could measure key statistics coaches wanted about their players. But that data was trapped at the individual-test level. There was no platform for coaches to see it across their whole roster, or to track how any one player was trending over time.

That gap is what I set out to close.

Pison had launched an app with baseball teams to test how our watch’s motion sensitivity could measure key statistics coaches wanted about their players. But that data was trapped at the individual-test level. There was no platform for coaches to see it across their whole roster, or to track how any one player was trending over time.

That gap is what I set out to close.

The Challenge

Our dashboard needed to surface cognitive data, a type of data most coaches had never encountered before, alongside the performance data they already knew how to read. That created four compounding problems:

No existing mental models
No existing mental models

Users had reference points for performance stats, but not for cognitive metrics like Readiness, Agility, and Focus.

Users had reference points for performance stats, but not for cognitive metrics like Readiness, Agility, and Focus.

Cognitive overload
Cognitive overload

A fully detailed visualization risked overwhelming coaches rather than helping them.

A fully detailed visualization risked overwhelming coaches rather than helping them.

So many possibilities
So many possibilities

Cognitive data could surface a huge range of insights, which meant the real design problem was narrowing down what to show, not what was technically possible to show.

Cognitive data could surface a huge range of insights, which meant the real design problem was narrowing down what to show, not what was technically possible to show.

Two views
Two views

Coaches needed to see both team-wide patterns and individual player trends over time, without those two needs fighting each other for space on the same screen.

Coaches needed to see both team-wide patterns and individual player trends over time, without those two needs fighting each other for space on the same screen.

Research

I approached this with three parallel, overlapping tracks: coach interviews, data visualization research, and iterative design reviewed with stakeholders, all running concurrently over about four months.

  1. Coach Interviews

  1. Coach Interviews

I talked with four coaches at the college and MLB level who were piloting our dashboard. We discussed the specific metrics they already relied on from competitor platforms, what those charts looked like, and how to meaningfully pair that familiar performance data with our newer cognitive metrics.

  1. Data Visualization Research

  1. Data Visualization Research

Alongside the coach conversations, I worked with our data engineers and CPO to figure out how to best depict cognitive data, specifically, what was meaningful to show at the individual-player level versus the team level, since those are genuinely different questions a coach might ask.

  1. Iterative Design

  1. Iterative Design

For roughly four months, I met multiple times a week with coaches and data engineers, sharing drafts and progress and gathering feedback. Iterations worked through chart interactions such as zooming, hover states, color coding, and information hierarchy, along with testing different chart types entirely, refining based on what that recurring feedback loop surfaced.

The Solution

The final dashboard is built around three cognitive tests, Readiness, Agility, and Focus, each visualized consistently so that learning to read one chart transfers directly to the other two.

Engagement view shows compliance and ranking across the roster at a glance: athletes ranked by test completion rate, color-coded against a team compliance target, alongside a breakdown of total tests taken over the selected time period and a per-athlete test count.

Results view goes deeper on performance itself. A result-variability bar shows each metric's average against its typical range (in-range, slower, or faster), color-coded green, yellow, or red for fast scanning. Below that, rolling-mean line charts for each of the three tests smooth out short-term fluctuations to reveal real trends over time, paired with distribution charts comparing the current period against the previous one, so a coach can see not just where a player is now, but whether the whole team's spread has shifted.

Team average scores table rolls everything up into a single sortable view, daily averages for each metric plus test count, ranked across the full roster, for coaches who want the team-level read rather than digging into individual charts.

Across all three views, color and layout stay consistent (readiness in blue, agility in yellow, focus in red/orange), which was a deliberate choice to reduce the relearning cost every time a coach switches between team and individual perspectives.

Engagement view shows compliance and ranking across the roster at a glance: athletes ranked by test completion rate, color-coded against a team compliance target, alongside a breakdown of total tests taken over the selected time period and a per-athlete test count.

Results view goes deeper on performance itself. A result-variability bar shows each metric’s average against its typical range (in-range, slower, or faster), color-coded green, yellow, or red for fast scanning. Below that, rolling-mean line charts for each of the three tests smooth out short-term fluctuations to reveal real trends over time, paired with distribution charts comparing the current period against the previous one, so a coach can see not just where a player is now, but whether the whole team’s spread has shifted.

Team average scores table rolls everything up into a single sortable view, daily averages for each metric plus test count, ranked across the full roster, for coaches who want the team-level read rather than digging into individual charts.

Across all three views, color and layout stay consistent (readiness in blue, agility in yellow, focus in red/orange), which was a deliberate choice to reduce the relearning cost every time a coach switches between team and individual perspectives.

Impact

Product launched under pilot program to refine product before wider release.

The dashboard is now live and in use with baseball teams at the high school, college, and professional level, currently piloting with 12 teams. Feedback so far has been positive, with coaches specifically calling out the need for help interpreting what the cognitive scores actually mean in context, not just seeing the numbers, but understanding what to do with them.

That feedback is now shaping the next phase of work: I'm partnering with our data engineering and ML teams to build interpretive guidance directly into the dashboard, so the platform doesn't just display cognitive data but helps coaches act on it.

The dashboard is now live and in use with baseball teams at the high school, college, and professional level, currently piloting with 12 teams. Feedback so far has been positive, with coaches specifically calling out the need for help interpreting what the cognitive scores actually mean in context, not just seeing the numbers, but understanding what to do with them.

That feedback is now shaping the next phase of work: I’m partnering with our data engineering and ML teams to build interpretive guidance directly into the dashboard, so the platform doesn’t just display cognitive data but helps coaches act on it.

Reflection

A key design challenge in this project was producing visualizations for data users had never seen before. Without established patterns to lean on, the interface needed to be highly intuitive, consistent color and chart logic across all three metrics, contextual range indicators, and a clear separation between team-level and individual-level views, so coaches could build confidence interpreting unfamiliar data without a manual.