Educational data analysis oriented towards decision making and improvement of educational processes, also known as Learning Analytics or Academic Analytics, has raised increasing interest among researchers and other educational agents and stakeholders. The idea behind Learning Analytics is that the large volumen of data (popularly known as big data) used in ICT-based educational processes is far too high to be useful for decision making without further processing. Therefore, application of adequate data mining, analysis and visualization techniques is required in order to better understand educational processes and support decision making for all agents involved: teachers, students, course designers and educational institutions. To this day, these techniques have focused on the relation between data recorded in e-learning systems, such as Learning Management Systems (LMS) or Virtual Learning Environments (VLEs). Nevertheless, the emergence of other ICT-based learning modalities, such as Massive Online Open Courses (MOOCs) or Personal Learning Environments (PLEs) have given new impulse to this field of study. The interest on Learning Analytics has gathered researchers from many different fields. Thus, scholars from very different areas, such as computing, information systems, education, psychology, statistics, semantic and network analysis, among others, are orienting their efforts towards the application of different techniques and approaches aimed to study and understand the information extracted from educational data, as well as the improvement of training and educational processes. Some examples of these improvements are focused on increase of students’ academic performance and achievement, the design of early-warning systems to predict and lower attrition rates, automation of adaptive educational systems, visualization of aggregated data, enhancement of social dynamics or the use of recommendation systems for training and education. However, current research is astoundingly disperse in terms of application field and scope. And this dispersion hinders in most cases generalizability of results. Thus, it is possible at this time to come across empirical studies on Learning Analytics in settings as small as a single course with few students, and analysis of a great number of courses with thousands of students in MOOCs or large organizations. Furthermore, determining the adequacy of one or another approach to the analysis of educational data becomes a difficult task. Hence, while analyses based on small datasets (small data) may offer very detailed information for specific settings, the analysis of big data may provide relevant information for wider scope course management. Even more, as of now there is no consensus on a valid criterion to determine the limit between small and big data. Because of the many different approaches to Learning Analytics, the main objective of this Workshop is to bring together the multiple perspectives and disciplines and establish a common framework of Learning Analytics practices and research experiences in this field. As a result of this approximation of different approaches, two main results are expected from the research contributions in this Workshop: first, the sharing of Learning Analytics initiatives as a first step towards fostering and strengthening collaboration among participants, aiming for a wider generalizability of results; and second, the comparison between research based on small data and big data may offer a space to debate and find common issues which facilitate future investigation of issues in Learning Analytics.