VA for Aircraft Safety & Reliability
This project is conducted in partnership with The Boeing Company, Aeroinfo, and MITACS. Our Integrated Science approach begins with work onsite at a company or agency to understand and support analysis and decision making in that organization through enhanced information technology. In many cases a barrier to effective implementation of new technology is a lack of knowledge of the effectiveness of that technology for a particular class of cognitive tasks that much be performed by that organization. This generates research questions for empirical studies of human cognition and collaboration in the organization and in our laboratory. The results of this work inform the selection and customization of technological systems for cognitive support, enhanced design of those systems, and methods for training users and integrating visual analytic support technology within an organization and across collaborating organizations.
The project is designed to assure integration of key aspects of visual analytics: technology design, onsite and distributed collaboration with Aeroinfo (Boeing Canada) customers, and direct support for core business operations at Aeroinfo. The proposed project expands our examination of current technological support for human reasoning about risk and vulnerability within AeroInfo, their parent The Boeing Company, and its airline customers. The internship and our overall program contribute directly to the methods Aeroinfo uses to improve the design and maintenance of Boeing aircraft. It will also support Aeroinfo to expand the scope of their business, improve operations, and in general maintain their competitive edge. Based upon these findings we will conduct labroatory studies focused on key aspects of analytic cognition. These will drive the design and testing of new interaction prototypes.
Our current proposal is built around "Pair analytics" projects with Aeroinfo clients at Boeing and airlines on aircraft maintenance analysis. These projects consist of a Visual Analytic Intern (VAI) working with a Aeroinfo/Boeing subject matter expert (SME). Together they examine a maintenance and/or safety dataset that contains information that can support their analysis. The results of this are analyzed in our laboratory to determine how best to build visual analytic technologies and analytic processes at Aeroinfo. Prior to each session problems must be identified by Aeroinfo and their clients that are important for the success of their business operations. One or more SMEs must be proposed and the intern must work with them to select the appropriate visual analytics tool and dataset. Those datasets must be cleared for access by our intern and preprocessed by them for use. Finally, the intern must gain sufficient proficiency in the use of the tool and in background information on the problem so as to collaborate with the SME in analysis. The analytic session is conducted, and the captured video and screen data are analyzed in our laboratory for usability and utility of the tool, analytic process effectiveness and efficiency, and ease of collaboration in the pair. The results of this process are fed into the development of new interfaces to visual analytics tools through an iterative prototyping and evaluation process that will inform future visual analytic pair sessions at Aeroinfo and their clients. Topics for these sessions have ranged from analysis of birdstrikes to the impact of Icelandic volcanic activities on the airline industry.
Representative fieldwork milestones are:
* Co-design of an aircraft safety-related analytic task with a domain expert. Sub-tasks: identification of analytical task, preparation of the dataset, selection of visual analytic tool. Deliverables: aircraft-safety analytical task and dataset
* Preliminary analysis and identification of outcomes Milestone: Outcomes of analysis and establishment of boundaries on the task to be used in evaluation of pair analyses
* Implementation of four Pair Analytic sessions on aircraft safety task/data. Deliverables: primary data to be analyzed: video-recording of the sessions, screen-capture of the sessions, results of the analyses
* Quantitative analysis using cognitive flow, insight generation and confirmation bias metrics Milestones: Operational concepts, coding scheme, coding, intercoding reliability, and statistical analysis and identification of critical tasks
* Task: Qualitative analysis to identify core variables that affect dyad-performance in Pair Analysis Milestones: grounded theory applied, operational concepts variables on performance of pair analysis
Video demos of analysis sessions:
Bird Strike Analysis using INSPIRE
Bird Strike Analysis using TABLEAU
Key Personnel
Andrew Wade
Andrew conducted these sessions as part of his MSc. in SIAT while he was an intern at The Boeing Company.
Dr. Richard Arias-Hernández, Ph.D.
Dr. Arias-Hernández coordinated Aeroinfo in Richmond BC and at Boeing, their main client and parent company in Seattle.