Project thumbnail image

College of Engineering Unit(s): 
Electrical Engineering and Computer Science

Team: 
Yujie Wang, Blake Le, Mateo Rey-Rosa and Zachary Kauffman

Project Description: 

Data visualization is also a kind of machine learning, which can accept user survey input, summarize the candidates' choices, and display all results and summaries in a convenient way.

The ability to efficiently assess the efficacy of a data visualization is at the core of this project. When a data visualization designer creates a visualization, they are not always certain that people are able to understand the data that is being portrayed. Sometimes data visualizations are very hard to comprehend, or lead readers or customers to the wrong conclusions. To get an idea of the efficacy of a data visualization, data visualization designers like to receive feedback on their designs. One of the ways that this can be done is by using a survey taking platform that allows a data visualization designer to pay for survey takers. However, the process for receiving feedback from data visualizations is a manual intensive process, which is in need of automation. Previously, a data visualization designer would have to go through the manual process of creating a survey through Qualtrics, exporting the survey, and uploading it to a service such as Amazon MTurk that would collect feedback from a specified sample of survey takers. The data from the feedback would then need to be analyzed to determine how effective the data visualization was in conveying information. A survey would therefore consist of questions related to identifying data points in the data visualization or asking questions that give the designer a good understanding of how many people are able to quickly understand the data being portrayed. This entire process is attainable when there is a small number of data visualizations that need to be processed, but fails to scale for a large number of data visualizations.

In order to automate the feedback collection process for data visualization, an automated process needs to be created. The goal of the project is to create a website where designers can upload the data visualization results of their survey results, and then automatically send them to the survey platform Amazon MTurk. Data visualization designers will be able to specify how many survey responses they want, as well as other criteria such as age or demographics. After the survey is completed, the data visualization designer will be notified and can then view the survey results. The data visualization designer can also check the progress of the survey, which indicates the need to interact with Amazon MTurk to retrieve the current progress every time the user wants to check. Finally, the data visualization designer will be able to perform a set of statistical tests on the returned data to understand and interpret the survey results. From here, data visualization designers can adjust their creation and make modifications to achieve more effective visualization. For the time being, the combination of the website and the server used to process the survey results will be the scope of the final project and will be provided as a service. The project is likely to be launched within OSU as a service that students or professors can use, but it can also be expanded to a service that can be used by the public.

Project Website(s): 

Project Communication Piece(s): 
AttachmentSize
PDF icon winter_term_retrospective.pdf226.19 KB
Project Communication Pieces do not open in a new window. Please use your browser's back button to return to this page.