Reading the data

In our efforts to deliver value to the user, we strive to do it in a way that is as much as possible grounded in evidence and data. Even so, users must be able to read the data and understand what they are seeing. At some point, the use of intrinsically scientific tools becomes unavoidable.

Simplifying complex data down to a single "value" or "index" helps reach a broader audience, but it comes at the cost of hiding crucial information about a topic that is, by nature, anything but simple. Trends and patterns, for example, are easily lost when data is reduced to simple indices.

Simplification is not the only method. Offering diverse perspectives on the same data, and explaining the tools used to generate those perspectives, can yield tremendous value by conveying the inherent complexity of the topic. This is an important aspect of our work. Since UX is one of the three pillars of the current PoC, we invest considerable effort in both implementing and communicating these different views.

The scenario below illustrates this principle.

Consider an air-quality sensor measuring PM2.5 and PM10 inside an apartment. Readings are taken at regular intervals, say every 5 seconds. Over a sufficiently long period, the sequence of points appears as continuous lines. The left-hand view of the figure below shows these lines for the month of April 2026 (red = PM2.5, blue = PM10). At this stage, a user has already received a substantial amount of information.

A different way of examining the same underlying data is by using a heatmap. The data is arranged on a grid in which each axis represents a distinct dimension. In the right-hand view of the figure below, the horizontal axis shows the day of the month, and the vertical axis shows the hour of the day. Each cell contains the maximum value recorded during that one-hour interval, colour-coded according to its intensity. The legend helps identify the WHO (World Health Organisation) daily reference thresholds, similar to the horizontal dashed lines in the left-hand view. This perspective adds different information and makes it easier to see, for example, whether air quality is consistently better during particular hours of the day.

Personalisation rests, among other things, on the availability of different perspectives. A user may find that one view presents more personally relevant information than another, even though both are derived from the same data. Several challenges arise in this area:

• What is the optimal number of perspectives to maintain a high-quality user experience?

• Each perspective necessarily carries only part of the information; how can the resulting bias be reduced?

• Every perspective relies on a specific "tool" or "artefact," and we cannot assume that users already know how to interpret it. What is the appropriate amount of educational content to embed in the UX while keeping the focus on the air quality itself?

This is precisely where UX and research intersect. As noted in "Visualising the concept" it is also an area we plan to explore further.

Images taken from the AirQ PoC app: line chart view on the left-hand side; heatmap view on the right-hand side