Abstract:

The modern world is inconceivable without information technologies. The biggest part of our work, entertainment, social communications and education takes place in Internet and is driven by smart management systems. Each activity and sentiment, each progress and challenge are tracked and stored in a database. BIG DATA are collected and the BIG question that arises is how to analyze and use them in a proper way to improve our life style.

This article is a part of a larger research that aims to explore how these data can assist education and improve learners’ achievements. It makes an overview of the existing methods for learning analytics (LA) and the corresponding tools for data visualization.

The first part of the paper focuses on the vast amount and various types of data gathered through learning management systems. A detailed review is made of the two main approaches used nowadays for collecting data on learners:
• clickstream in online educational platforms estimating online trainee’s behavior and
• natural language processing (NLP) for sentiment analysis evaluating student‘s participation in forums and social communications, posting comments and replies, sharing emotions.
Practical cases, studies and results are discussed.

The second part of this paper aims to reveal the power of data visualization. A wide range of visual components like charts, histograms, gauges, word clouds, concept maps, heat-maps, tree-maps and many more are described with their pros and cons and are illustrated in real educational use.

Software tools for effective data analysis and visualization are discussed for both approaches: clickstream dashboards and skill meters as well as open source tools for NLP like:
• Writing Assessment Tool for evaluating quality of writing,
• Tool for the Automatic Analysis of Lexical Sophistication for lexical complexity,
• Tool for the Automatic Analysis of Cohesion for logical consistency of text,
• Reader Bench for text mining and social network analysis and
• Sentiment Analysis and Cognition Engine for sentiment, social cognition and social order.

The paper gives and overview of contemporary software products for LA - from small add-ins or plug-ins for existing educational environments like Gizmo for Moodle to big visual analytics system like VisMOOC showing complex relations between video clickstream data from MOOC course and students’ achievements. LEA’s BOX web platform combining clickstream and NLP analysis on data derived from different sources and providing rich set of visual components is also conferred.

Recent works on LA tools for data collection and representation interoperability and standardization are discussed.

Finally, authors’ experience, using some of the presented methods and tools, is shared, the challenges are discussed, and recommendations for performance tips are suggested.

Keywords:

Learning analytics, visualization, analytics methods.