Despite the exponential growth of the research on Learning Analytics (LA) and Educational Data Mining (EDM) over the last few years, the work has been still distant from the core Learning Sciences methods, theoretical constructs, and literature. At the same time, over the last 15 years, Learning Sciences as a field has been quite innovative, eclectic, and effective in incorporating new methodological stances, such as micro-genetic methods, micro-ethnographies, and design-based research. It seems that the time has come to build sound connections between these traditions. The goal of this symposium is to bring together researchers coming from different academic perspectives, to explore and examine common LA/EDM methodological and theoretical threads with wide applicability within the Learning Sciences. The papers presented explore text mining in clinical interviews, moment-by-moment learning curves and traces, data mining of programming logs, and cognitive tutors, representing the main perspectives and methodological approaches in the field.