Exploring the Progression of Early Programmers in a Set of Computational Thinking Challenges via Clickstream Analysis

In recent years, many initiatives have aimed to develop basic computational thinking skills. Despite the popularity of online platforms for early programmers, we still lack detailed information to analyze how these skills are acquired. In the present study we analyzed clickstream data from 3,355 participants enrolled in several Computational Thinking workshops using Kodetu, an online platform with fine grained logging features. Participants used Kodetu’s coding blocks to solve challenges of increasing difficulty while we gathered their clickstream in the platform. Here, we present our findings after evaluating these data in regards of participants’ characteristics (age, sex, previous knowledge), similarity with previously submitted solutions, and degree of discrepancy from the optimal solution. To facilitate collaboration with other researchers in this area, we released our dataset under an open license. To the best of our knowledge, this is the largest Computational Thinking-related datasets publicly available.