# Uncover gaze behaviour on phone screens with Neon

Products used: NeonCloud

Have you ever wondered what your eyes focus on when scrolling through your favourite app?

In this article, we test whether Neon, our latest eye tracker, can accurately capture and characterise viewing behaviour while gazing at small icons on mobile applications. To achieve this, we used some of Alpha Lab's existing tutorials, including how to generate scanpaths, define areas of interest (AOIs) and calculate outcome metrics, and map gaze onto dynamic content.

Shall we start?

# What you'll need

Below you can find the tools we used for this project. Using these, you can replicate the content of this article with your own applications.

# Cloud enrichment

# Alpha Lab tutorials

# How we used them

We first used the Reference Image Mapper enrichment to create a 3D model of a phone positioned on a desk and subsequently to map gaze onto a 2D image of the phone-on-desk. Then we used Alpha Lab tutorials to process the exported results, generate advanced visualisations, and compute outcome metrics.

💡 When preparing the Reference Image Mapper Enrichment, make sure your phone is stable on a phone mount or stand. The scanning video needed for this tool requires relatively static features in the environment. If there is a lot of movement or the objects change in appearance or shape, the mapping can fail. More on this in the docs!

# Gaze behaviour on mobile apps: Insights from Neon

# Heatmaps and scanpaths

What catches your attention and how do you visually navigate through the interface/page of a mobile app?

Two visualisations that help to illustrate these patterns are heatmaps and scanpaths (left and right panel below). Heatmaps show the areas of the app that receive the most attention, with warmer colours indicating more fixations, and can be generated natively in Pupil Cloud. Meanwhile, scanpaths trace the path of the eye movements, showing the sequence of fixations and eye movements that occur during the visual exploration. The circle size of the scanpath below reflects fixation duration: The bigger the circle, the longer the user fixated on this area of the screen. Use the scanpath tutorial linked above to generate this visualisation.

# Calculation of gaze metrics on AOIs

Analysing eye tracking data can provide valuable insights into user behaviour, but simply looking at visualisations like heatmaps and scanpaths may not always reveal the full story. This is why we opted for a quantitative analysis of our data as well by calculating gaze metrics, such as dwell time and time to first contact. These metrics offer tangible and quantitative outcomes about the salience of each AOI: Longer dwell time implies longer total fixation on a specific AOI and could be considered as a proxy of attentional allocation. Conversely, the shorter the time to first contact, the faster this AOI captured the user's attention, pointing to increased salience of this area. Follow along with the AOI tutorial for these calculations and charts!

# Map your gaze onto dynamic phone screen content

So far, we've only scratched the surface by examining static images. Now, it's time to dive into the dynamic world of our smartphones and explore gaze behaviour more naturally while scrolling. Use the dynamic screen mapping tutorial for this one!

Checking out the recording, Neon's accuracy richly captures gaze behaviour and provides a nice high-level overview of what the wearer was looking at.

Visualisations are great, but the real power of this tool is that it generates a CSV file containing gaze data mapped onto the screen, in 2D x, y coordinates. This offers many possibilities for further customisation and in-depth analysis.

# In the wild

That's all fine, but what does it look like to interact with phone screens when out and about? Consider this scenario: Imagine searching for products on your mobile app while looking at the physical products on a supermarket shelf. We brought Neon into the wild to assess its performance beyond controlled environments. See an example of real-world user behaviour below!

# Let's wrap it up!

This article has unveiled the remarkable capability of Neon's calibration-free eye tracking to capture viewing behaviour during mobile app interactions - both inside and out and about. This tutorial, and its outcomes, are not limited to our specific use case and could be particularly useful for other types of UI/UX research. By combining Neon with the techniques we've highlighted here, you can gain invaluable insights into user engagement.

Curious about how Neon can fit into your work? Need assistance in implementing your own analysis pipelines? Reach out to us by email or visit our Support Page!