Hawkeye makes it easy to export eye tracking data collected during tests. This lets you dig into the numbers and analyze your data however you'd like.

Hawkeye separates data into different files for each webpage or image a test subject interacts with. For example, if the user starts on your homepage and then visits their profile, the data will be separated out into two CSVs.

Hawkeye exports include three types of data: Fixations / Eye Positions, Touches, and UI Element Stats. You can view a sample export here.

Fixations / Eye Positions

During a test, Hawkeye captures where on the screen the user is looking 30 times per second. A list of these eye positions is available in the raw-eye-position-data folder of your export.

For each eye position, the following information is included:

  • started_at: The time the eye position began. Generally, this will be the same as ended_at.
  • ended_at: The time the eye position ended. Generally, this will be the same as started_at.
  • element_uid: The UID of the UI element the user was looking at.
  • standard_deviation: An estimation of how much the user's eyes were moving at this point in time. Equals the standard deviation of the 10 previous eye positions
  • x_pos: The x position on the screen the user was looking at, adjusted for scroll offset. Measured from the left side of the screen.
  • y_pos: The y position on the screen the user was looking at, adjusted for scroll offset.
  • abs_x_pos: The absolute x position on the screen the user was looking at, without taking into account scroll offsets.
  • abs_y_pos: The absolute y position on the screen the user was looking at, without taking into account scroll offsets.
  • relative_x_pos: The x position of where the user was looking within the nearest UI element.
  • relative_x_pos: The y position of where the user was looking within the nearest UI element.

All screen positions are measured from the top left corner of the screen with x representing the horizontal axis and y the vertical axis.

The difference between position, absolute position, and relative position

A common point of confusion is the difference between the three different positions included for each eye position.

Pos equals the eye position adjusted for the scroll offset. For instance, if the user scrolls down the page, y_pos will be the distance from where they are looking to the top of the scroll content, not to the top of the phone.

Abs_pos equals the eye position relative to the screen, without taking into account scroll position. For instance, if the user looks at the same position on the screen while scrolling, abs_pos will stay the same.

Relative_pos equals the eye position relative to the UI element the user is looking at. For instance, if the user is looking at a button, relative_pos tells you where on the button they are looking taking the top left of the UI element as zero. This is useful if you want to track looks within a particular UI element.

Touches

Hawkeye also records the taps and scrolls a user performs during a test. A list of each location the user touches the screen is available in the raw-touch-data folder of your export.

For each touch, the following values are recorded:

  • x_pos: The x location of the touch, adjusted for scroll position.
  • y_pos: The y location of the touch, adjusted for scroll position.
  • element_uid: The identifier of the element the user is touching.
  • timestamp: When the touch occurred, relative to the start of the test.
  • relative_x_pos: Location of the touch in the x direction relative to the element being touched.
  • relative_y_pos: Location of the touch in the y direction relative to the element being touched.

UI Element Stats

In addition to raw eye position and touch data, Hawkeye automatically generates statistics for each UI element the user looks at with. A list of these stats can be found in the ui-element-data folder of your export.

The following information is included for each UI element:

  • uid: The UID of the element
  • looks_count: The number of times the user looked at this element.
  • unique_looks_count: The number of users who looked at this element at least once.
  • avg_first_look: The time the user first looked at the element.
  • avg_last_look: The time the user last looked at the element.
  • avg_look_duration: The average duration of a fixation on this element.
  • taps_count: The number of times the user tapped this element.
  • unique_taps_count: The number of users who tapped this element at least once.
  • avg_first_tap: The time the user first tapped the element.
  • avg_last_tap: The time the user last tapped the element.

Did this answer your question?