Reproducible Workflows for Exploring and Modeling EMA Data

Improper use of substances like cannabis may lead to physical, emotional, economic, and social problems. Therefore, it is significant to elucidate the inter-individual and intra-individual influences along with contextual influences that predict the use of cannabis. TigerAware is a mobile survey data collection platform that holds unique promise to advance research in addiction and substance use. This paper presents a novel method to support Ecological Momentary Assessment (EMA) studies. We propose to extract useful information from TigerAware survey data using data mining and machine learning methods, and structure customizable survey analyses into reproducible workflows. Through our analysis pipeline for EMA, researchers are able to discover meaningful information from survey data with minimal duplication of effort and improve the efficiency and rigor of the process.