ELIA Dataset: 70 Million Learning Logs from 20 Months of Web Browsing and Informal Vocabulary Learning

This dataset contains 70 million timestamped learning events automatically logged for nearly 15,000 English learners over a 20-month period while browsing authentic English-language websites using ELIA, a browser plugin that supports vocabulary learning through word lookup, lexical recommendations, and retrieval practice. Collected between June 2023 and January 2025, the dataset captures rich, real-time interactions between learners and the ELIA edtech app, covering both on-page behaviours and practice activities. The collection offers a large-scale, longitudinal, and ecologically valid record of informal digital learning of English. It enables research on learner behaviour, strategy use, engagement patterns, lexical development, and learning outcomes in authentic online environments.