Fall 2022 Project Day Highlights Capstone and Master’s Projects

The first in-person Project Day was held at the end of the Fall semester to showcase completed undergraduate capstone projects and master’s software development projects. The event also celebrated the successes of ICS competitors in the Hawaii Annual Code Challenge (see previous post) and international Meteor Hackathon (see previous post).

Undergraduate Capstone Projects

Hawaiian Airlines Web Modernization

Presenter: Kobey Arai
Sponsor: Rajesh Nimmalagadda, Manager IT Digital Experience, Hawaiian Airlines

The Hawaiian Airlines website is one of the primary ways guests can book and manage their flights and view details about the state of Hawaii. This project involved enhancing multiple features of the Hawaiian Airlines website. These features include enhancing the functionality that guests utilize to book a ticket, from flight search to review-and-payment. Additionally, improvements were also made to the functionality that lets guests view and update details about their upcoming itinerary(s) and chat with customer representatives.

NLP for Aircraft Maintenance

Presenters: Guanhong Li, Nicholas Whiteley, Karen Wong
Sponsor: Lt. Col. Thomas Kline, U.S. Marine Corps Forces Pacific

In this engagement, ICS students worked with the U.S. Marine Corps Forces, Pacific, in utilizing natural language processing (NLP) techniques to improve aircraft maintenance activities. Currently, Work Unit Codes (“WUCs”) must be manually chosen by traversing a large tree, making some inputs erroneous or too broad. Existing processes are time-consuming to acclimate to and regularly use. As part of the solution, the team implemented a Python web application using natural language processing on historical data and WUC trees to convert user inputs into WUC’s. The application takes in user input description, then utilizes NLP to predict possible matching WUC’s, and finally displays the likelihood percentage along with the corresponding description for each WUC.

Fairy Dust Fighter Logistics Wargame

Presenters: Hansen Cabanero, Michael Hui, Tony Long
Sponsor: Lt. Col. Thomas Kline, U.S. Marine Corps Forces Pacific

The objective of this U.S. Marine Corps Forces, Pacific project is to create a homegrown logistics simulation based on executable, sustainable, and real-world constraints. The Power Business Intelligence and Python-based AI solution enable the simulation of concepts such as running out of fuel, water, ammunition, or medical supplies, critical cargo shipments needed to fix a downed transport aircraft, enemy engagement, medivacs, etc.

Phase Maintenance Inspection Software for Rotary Wing Aircraft

Presenters: Jonathan Ma, Lise Marie Nilsen, Adam Joseph Parrilla, Kirsten-Elise Rensa
Sponsor: Major Curtis Bew, U.S. Army Commander, B/209th ASB, 25th Combat Aviation Brigade Hawai‘i

To improve the scheduling of aircraft maintenance of the U.S. Army’s 25th Combat Aviation Brigade, the ICS student team built a web application that provides a basic interface for a Phase Maintenance Inspection Scheduler to organize and manage military rotary wing aircraft as it goes through its maintenance inspection.

Scientific Workflow Visualization

Presenters: Kristian Lozo, Yeiji Han
Sponsor: Dr. Henri Casanova

The main goal of our project is to implement a tool to visualize workflow structures as graph vertices and edges in a level-by-level fashion in the browser. WfCommon workflow description visualizes the workflow structure by implementing schema validation to take a JSON file as input. To visualize the workflow structure, we use Cytoscape, an open-source graph visualization library. Each Cytoscape vertex must be mapped to a workflow task, and edges correspond to workflow parent-children relationships. We implement a recursive algorithm for each level of a task in the workflow. Further, we design and evaluate greedy heuristics to reduce the number of edge crossings in the workflow. In addition to that, implementing a random heuristic that greedily swaps tasks within the level is to reduce the number of edges crossing more.

tsDetect Enhancement Reporting

Presenters: Feimei Chen, Michelle Leano
Sponsor: Dr. Anthony Peruma

This engagement involved making enhancements to an existing open-source test smells detection IDE plugin – tsDetect. The major enhancement involved the implementation of an “Analysis Summary” UI that provides a summary of the test smell detection analysis. This summary includes a Smell Type Summary that shows the change in the count of

detected smells and the most commonly occurring smell type, a File Summary that shows the count of files analyzed, those that have smells, and the file with the most number of smells, and finally, a Method Summary that shows the change in the count of test methods that exhibit test smells and the smelliest method.