Workshop: Sharing and Reusing Data and Analytic Methods with LearnSphere

Monday, March 5, 2018 (Full-Day Workshop) at LAK 2018 in Sydney, Australia

Organizers

John Stamper, Carnegie Mellon University
Kenneth Koedinger, Carnegie Mellon University
Carolyn Rose, Carnegie Mellon University
Philip Pavlik, University of Memphis

Workflow How-To Videos

LearnSphere Workflow Tools: Overview

BKT Analysis using LearnSphere Workflows

Creating new workflow components

Questions?

Contact John Stamper: jstamper AT cmu DOT edu

This workshop will explore LearnSphere, an NSF-funded, community-based repository that facilitates sharing of educational data and analytic methods. The ultimate goal is to create the building blocks to allow groups of researchers to integrate their data with other researchers in order to advance the learning sciences as harnessing and sharing big data has done for other fields.

In the workshop, we will focus on Tigris, a workflow tool within LearnSphere that helps researchers share analytic methods and computational models. Authors of accepted workshop papers will integrate their analytic methods or models into LearnSphere’s Tigris in advance of the workshop, and workshop attendees will have drag-and-drop access to these analytic methods through Tigris. We will learn about these different analytic methods during the workshop and spend hands-on time applying them to a variety of educational datasets available in LearnSphere’s DataShop. Attendees can also bring their own datasets to the workshop. Finally, we will discuss the bottlenecks that remain, and brainstorm potential solutions, in openly sharing analytic methods through a central infrastructure like LearnSphere.

Submission Information. We invite submissions of short papers that describe analytic methods that authors are willing to share within LearnSphere's Tigris workflow tool. Papers should be 2-4 pages long and describe the inputs, algorithm/model, and outputs of the proposed analytic method (you can use this submission template). They can include analytic methods that have previously been published on. The goal of this paper is to describe the mechanisms in enough detail that we can integrate the code into Tigris.