Welcome

The Project

ELAT is a web-based tool to pre-process the edX metadata and logfiles into a structured database, to generate csv files for analysis and visualizations. Given that each edX logfile contain the interactions of all learners for all the courses of an institution throughout a day, filtering the useful information for a researcher or instructor of a particular course may be time consuming and challenging. Here is where ELAT shines, as it handles all the pre-processing of the logfiles and generates structured data that can be used for further analysis, such as experiments in R or Python. The process is fully executed in client side (on the browser) to minimize privacy and security risks. It is developed in the Delft University of Technology, based on the edx-analysis approach, read more about it here.

Who is ELAT for?

Researchers, lecturers, educators, and anyone interested in analyzing and experimenting on the edX platform, using the information buried in logfiles. It requires minimal effort, no installation, no programming experience.

Important Details/Limitations

  • Currently ELAT can only store a single course at a time, but the database can be easily deleted and then other courses can be analysed
  • Currently ELAT is only developed and tested in Google Chrome
  • The minimum duration of a session (from first to last event) is 5 seconds. Series of events with less than 5 seconds between start and finish are discarded. (ELAT considers a “session” as a series of events of a certain type for a single user.)
  • Sessions end when the student closes the course element they were interacting with, or after a period of inactivity. Currently, this period is 30 minutes.
  • Because of processing limitations, currently ELAT cannot retain sessions that start one day and end in the next, since they belong to different files and the files are processed one by one. This means that some overnight sessions are split into two separate sessions
  • Because of file size limitations (files over ~73MB compressed, ~2GB uncompressed), ELAT has to chunk some of the bigger into smaller pieces. It does not happen often, but when it does, the sessions that start in a chunk and continue to the next are split in two, like overnight sessions.

Current Status

  • Running everything on client side, no server, no information sent over the internet.
  • Deployed on GitHub Pages
  • Working IndexedDB database, schema adapted from the MOOCdb Model
  • Decompress and process gzip files in JavaScript
  • Take contents from decompressed log file, process in-browser, then store into IndexedDB
  • Generate main indicators and graphs
  • Prepare csv files for download from the tables in the database