Analyzing Learning Patterns.


Data insights from Curious
Learning's deployment of
tablets around the world




The increasing ubiquity of smart devices, even in under-resourced communities worldwide, opens the path to innovative approaches for how to reach children with quality learning experiences. In this context, we have developed and deployed a mobile literacy platform with global reach for functionally non-literate populations, as well as to serve at-risk children in low SES communities in developed countries. One key challenge is to create effective educational experiences that support child-driven learning scenarios in contexts where children either cannot attend school, do not have access to adequately trained teachers, or have limited learning resources in their homes and need to reinforce what they learn outside of school. Given this dramatic diversity of learning contexts, an effective learning solution must be able to adapt in a data-driven manner.

Curious Learning is a novel technology platform and approach to addressing the challenge of literacy education using cloud-connected mobile devices with remote data monitoring and gathering. The content strategy is informed by neurocognitive principles for how the brain learns to read, and an educational strategy for child-driven learning informed by principles of social constructivist learning [8]. This platform is presently deployed in several countries including Ethiopia, Uganda, India, South Africa, Peru and rural United States. It serves approximately 5000 children and spans under-resourced learning contexts from children with no access to schools, children in schools with 50 to 100:1 student- teacher ratio, and in at-risk, under-resourced communities in developed countries.

In this project, a cross-cultural analysis of child-driven exploratory and learning behavior using a mobile literacy intervention in two radically different deployments was explored. The deployment data utilized for analysis includes that of children in a remote village in Ethiopia as well as a a deployment with at-risk children living in a rural town in the United States. The analysis offers early insights into how children explore the platform at different points in time; how they develop preference among applications; and how they interact with the platform’s user interface. These results can be useful to inform the development of adaptive user interface designs and educational apps that best utilize children’s innate exploratory and problem-solving behavior to support literacy or other learning-oriented tasks in these relatively poorly understood child-driven learning contexts.


  • data cleaning, organization and structure 
  • extensive analysis of learning behaviors in data using Python 
  • development of heuristics to capture learning patterns
  • visualization of data to capture and communicate differences in patterns of both populations


Python | SQL | Pandas

Screen Shot 2018-01-17 at 1.34.02 PM.png