- Computational Social Science
- Critical Data Studies
- Data Science
- Economics and Information
- Education Technology
- Ethics, Law and Policy
- Human-Computer Interaction
- Human-Robot Interaction
- Incentives and Computation
- Infrastructure Studies
- Interface Design and Ubiquitous Computing
- Natural Language Processing
- Network Science
- Social Computing and Computer-supported Cooperative Work
- Technology and Equity
Understanding Machine Learning with D3: Visualization for Models and Algorithms
Statistical machine learning is an increasingly valuable skill in today's data-rich world.
Traditionally machine learning papers and tutorials take a highly abstract, mathematical route to explaining models and algorithms. This mathematical focus not only creates barriers to entry, but can also create excessive confidence in models, whose theoretical guarantees may not survive contact with messy real-world data. Visual explanations through pictures and diagrams are often our most effective teaching tools in such situations.
In this talk David will show models and algorithms that lend themselves to visual explanations. From visualizing the models themselves, to their iterative steps, results and checking methodologies, this talk will demonstrate core machine learning concepts using d3.js.
This is a practice talk for OpenVis Conf 2014.
Staccato social support in mobile health applications
Social support plays an important role in health systems. While significant work has explored the role of social sup- port in CMC environments, less analysis has considered social support in mobile health systems. This paper describes socially supportive messages in VERA, a mobile application for sharing health decisions and behaviors. The short and bursty interactions in social awareness streams [Naaman] afford a particular style of social support, for which we offer the label staccato social support. Results indicate that, in comparison to previous work, staccato social support is characterized by a greater prevalence of esteem support, which builds respect and confidence. We further note the presence of ‘following up’, a positive behavior that contributes to supportive interactions, likely via social pressure and accountability. These findings suggest design recommendations to developers of mobile social support systems and contribute to understanding technologically mediated social support for health.
This is a practice talk for CHI 2014.