Let's look at some stories that a combination of data analysis and observations can reveal. The play tree below shows an example of a child building a word (her name) backwards. A look at the video data shows her being puzzled why the app doesn't say her name right. The facilitators then explained the concept of word direction to her, and she was able to correct her word. This is an example of a learning moment revealed by a bird-eye view of play sessions.

A sequence of images below is the timelapse of a scene created by a child playing with SpeechBlocks II app. They are reconstructed from the logs. Records from the classroom tell us two stories: (1) how her ideas were partially initiated by her, partially prompted by the app itself, and (2) how she used various input systems to make the words she wanted:

This is an example of how qualitative observations and log data can be combined to tell a richer story.

Sometimes quantitative analysis can complement the picture drawn from qualitative observations. In the analysis of play with SpeechBlocks II app, we noticed that children with lower self-regulation (specifically, executive functioning, EF) and literacy skills (specifically, phonological awareness, or PA) were likely to engage in distracted behaviors, such as random tapping and swiping and "taking pictures" using the word recognition interface. We looked at a few quantitative measures that we thought to be plausibly associated with these behaviors, such as number of touches per session (to capture quick tapping) and speed and "jerkiness" of finger movements (to capture quick random swipes). We saw that, indeed, all of them increased with low PA and EF, and sometimes the association was strong enough to be statistically significant. It corroborated our impression that the app was not sufficiently supporing the children with lower self-regulation and literacy skills. You can find more about this study in our Computers & Education paper.