Rebuilding the structure behind a growing dashboard

Rebuilding the structure behind a growing dashboard

Dashboard UX

Data Visualization

I restructured the IODA dashboard to give a growing platform the system behind it — reorganising how connectivity data is presented, how users interact with it, and how new information can be added without breaking what's already there.

IODA is a near real-time platform for monitoring internet disruptions around the world. It aggregates multiple data sources and surfaces them through interactive charts, helping researchers and analysts understand when and where outages occur.

Rebuilding the structure behind a growing dashboard

Rebuilding the structure behind a growing dashboard

Dashboard UX

Data Visualization

I restructured the IODA dashboard to give a growing platform the system behind it — reorganising how connectivity data is presented, how users interact with it, and how new information can be added without breaking what's already there.

IODA is a near real-time platform for monitoring internet disruptions around the world. It aggregates multiple data sources and surfaces them through interactive charts, helping researchers and analysts understand when and where outages occur.

THE PROBLEM

Growth introduced complexity without a clear system

New charts and signals had been added without a consistent placement logic.

Views became progressively denser and harder to interpret. Key interactions relied on desktop patterns that worked poorly in practice. The system no longer had a clear model behind it.

INFORMATION ARCHITECTURE

Structuring the dashboard around how the data works

The Country View had become overloaded, but not simply because there were too many charts. The deeper issue was that these charts represented different levels of analysis. Some belonged at the national level, some at the regional level, and others only made sense in the context of individual networks. By placing them all within the same view, the interface collapsed those distinctions and forced users to construct the system’s logic for themselves.

I addressed this by reorganising IODA into three distinct views:

  • National — high-level, aggregated signals

  • Regional — geographic breakdown

  • Network — ASN-level detail

This was architectural work as much as interface work. By separating charts according to the level at which they were meant to be read, I strengthened the mental model of the product itself. Each view became easier to scan, easier to interpret, and better able to absorb future signals without collapsing back into clutter.

The Country View had become overloaded, but not simply because there were too many charts. The deeper issue was that these charts represented different levels of analysis. Some belonged at the national level, some at the regional level, and others only made sense in the context of individual networks. By placing them all within the same view, the interface collapsed those distinctions and forced users to construct the system’s logic for themselves.

I addressed this by reorganising IODA into three distinct views:

  • National — high-level, aggregated signals

  • Regional — geographic breakdown

  • Network — ASN-level detail

This was architectural work as much as interface work. By separating charts according to the level at which they were meant to be read, I strengthened the mental model of the product itself. Each view became easier to scan, easier to interpret, and better able to absorb future signals without collapsing back into clutter.

UI DESIGN

Making data relationships visible and interpretable

The signal selector was redesigned so that relationships between data sources are explicit rather than implied.

Colour-coded labels Selected signals now appear as colour-coded labels tied directly to their corresponding chart lines. This reduces the mental work required to map controls to what's shown on screen.

Multi-select Supports comparison without losing context. Default selections preserve familiar behaviour for existing users.

View switcher A toggle was added to let users switch between a stacked and overlayed chart view. Users who want to compare signals side by side can do so, while those looking at trends over time can switch without losing their place.

Chart Description Each chart now includes a short description directly beneath its title. This reduces reliance on hover and communicates purpose at first glance, particularly for users encountering a signal for the first time.

METRIC PLACEMENT

Placed by meaning, not just availability

Packet loss appears at the national level, where its aggregated nature makes it suitable for high-level insight.


Upstream delay is shown at the network level, where its precision aligns with ASN-level analysis.

This reinforces the meaning of each view and avoids misleading comparisons across the wrong level of detail.

Each new metric was placed according to how it should be interpreted.


Packet loss appears at the national level, where its aggregated nature makes it suitable for high-level insight.


Upstream delay is shown at the network level, where its precision aligns with ASN-level analysis.


This reinforces the meaning of each view and avoids misleading comparisons across the wrong level of detail.

Each new metric was placed according to how it should be interpreted.

Packet loss appears at the national level, where its aggregated nature makes it suitable for high-level insight.

Upstream delay is shown at the network level, where its precision aligns with ASN-level analysis.

This reinforces the meaning of each view and avoids misleading comparisons across the wrong level of detail.

OUTCOME

A system that can continue to grow without breaking

Connectivity data is now organised into clear, logical views

  • Users can compare signals without losing context

  • New metrics can be placed without disrupting the interface

  • Charts communicate their purpose without requiring interaction to understand

The dashboard now has a model behind it. Instead of accumulating features, it can continue to grow without becoming harder to use.