From CS294-10 Visualization Fa08
- Kuang Chen
- Heather Dolan
Since the initial presentation, we have more clearly refined the scope of the project. We plan to automatically generate forms for two datasets. The form generation algorithm will specify a static question ordering that respects topical constraints while following the survey design guideline of asking hard questions first. With this base assumption, we will create dynamic forms that respond to a user's input for a value with visual indications of where that value lies in the range of legal and expected values.
The code to generate an XML form specification, generating Adobe Flex forms from that specification, and learning of a statistical graphical model is done, as well as preparing data from two datasets/forms. A couple of "dynamic" data entry widgets are done too. We will demonstrate these aspects to the class.
In progress is the statistical model inference to generate an optimal question ordering, as well as inference to predict the conditionally expected form values for a field. We are also currently integrating the dynamic widgets to the generated Flex forms.
We have not started working the form layout algorithm, or more advanced dynamic widgets yet.
Initial Project Proposal
Data quality is very important. The point of entry is the ideal place to address data quality. This project proposes a design for an automatic form generation system that takes advantage of visualization techniques including widget choice and decoration. Automation organization of form questions is also an objective of the system. The forms produced using the system should help ensure efficient collection of high quality data. Additionally, as part of a larger project, a system for for users to provide comments and feedback on the form is also proposed.
The portion of the project for this class will focus on developing an algorithm for dynamic layout and pagination and widget design using helpful "decorations" that provide information about prior values or "risk" of a current value. Additional details can be found in the attached slides.
Proposed Division of Labor
Generate form from XML - Kuang
Design Decorations - Heather
Build decorations - Heather
Design Layout Algorithm - Kuang and Heather
Implement Algorithm - Kuang and Heather
Visualize Tanzania AIDS treatment forms - Kuang
Simple Evaluation - Heather
Write-up - Kuang and Heather
- Scented Widgets: Improving Navigation Cues with Embedded Visualizations. Wesley Willett, Jeffrey Heer, Maneesh Agrawala. InfoVis 2007.
- Guidelines for Designing User Interface Software. Sidney L Smith, Jane N Mosier. Technical Report
- Maintaining data integrity in a rural clinical trial. J Van Den Broeck, M Mackay, N Mpontshane, A Kany Kany Luabeya, M Chhagan, M. L Bennish. Clinical Trials Journal.
- A tool for the design of clinical forms supporting end-user integration. Georg Duftschmid, Thomas Wrba. Medical informatics and the Internet in medicine.
- EcoPod: a mobile tool for community based biodiversity collection building. YuanYuan Yu, Stamberger, Manoharan, Paepcke. JCDL 2006.
- Handheld computers for data entry. Tania M Shelby-James, Amy P Abernethy, Andrew Mcalindon, David C Currow. Trials.
- Automatically Generating Personalized User Interfaces. Krzysztof Z. Gajos. PhD Thesis.
- Online survey design guide. http://lap.umd.edu/survey_design/index.html
- Crowdsourcing user studies with Mechanical Turk. . CHI 2008.
- Using GOMS for user interface design and evaluation: Which technique? John, Bonnie and Kieras, David E. ACM Transactions on Computer-Human Interaction 3,4 (December 1996a), 287-319.
- Fitts’ Law: Modeling Movement Time in HCI http://www.cs.umd.edu/class/fall2002/cmsc838s/tichi/fitts.html
- Using the Keystroke-Level Model to Estimate Execution Times. David Kieras. ftp://www.eecs.umich.edu/people/kieras/GOMS/KLM.pdf
- Keystroke-level model for advanced mobile phone interaction. Holleis et al. CHI 2007