Streamlining a Decision Making Process in GeoSpatial Intelligence; Analyzing the Use of Artificial Intelligence Models in Healthcare and Patient Treatment Plans

Technical Project Abstract
A government contracting firm in Washington D.C needed an efficient decision making tool for Geospatial analysts overwhelmed by multiple predictive models. My team led the development of a web application that consolidates key information on each model, such as geospatial and temporal accuracy, usability, accessibility, and more. This enhanced collaboration and decision making efficiency. The backend was coded using Python and JavaScript with Django, and React was used for the frontend. It featured data visualizations built with D3.js and Chart.js to display user-friendly graphs from large datasets. Overall, the application was measured and predicted to save analysts up to 9 hours a week, and increased confidence in predictive analysis by 73%. After the initial production of the application, next steps would include scaling the application to ensure larger datasets can be used, testing for bugs found in our UI/UX evaluation, as well as incorporating more features.

STS Project Abstract
Artificial Intelligence (AI) has been making its way into many industries today, ranging from banking, healthcare, education, the federal government, and much more. With this newfound industry becoming more and more relevant in our daily lives, it is important to recognize the balance there must be between AI and human judgment, especially in fields that are human centric such as healthcare. Successful applications of AI in healthcare include image analysis, as well as diagnosis assistance. However a more controversial application of AI in healthcare includes IBM’s Watson for Oncology (WFO), which was once hailed as an impressive AI powered computing model to help patients get personalized cancer treatment plans. Many studies have concluded significant performance flaws in WFO, largely deviating from the opinions of multidisciplinary healthcare professionals when it comes to treatment recommendations. My research will analyze the ethical implications on integrating such a radical AI concept into a human focused field, like healthcare, highlighting the challenges in finding a balance between the two fields. In addition, I discuss the results of my research along with a literature review, showing the potential uses that AI has in healthcare as well as its faults and improvement areas. After presenting the results of my research, I analyze impact of the network created using Actor Network Theory (ANT) to demonstrate how AI affects all components of the healthcare system. I then conclude my discussion on the current priorities in AI integrated healthcare and how to best proceed with this integration of fields.

Connection
Throughout my summer internship, I created a web application that helped streamline the decision making process for choosing predictive tools in Geospatial analysis. This experience allowed me to see all of the different types of predictive models and analysis that our government uses, including AI integrated models. In my STS research, I was able to analyze the impacts of AI in healthcare, specifically in patient treatment plans. Similar to howWFO employs predictive analytics to enhance cancer treatment decisions, my summer internship work for streamlining geospatial intelligence decisions by employing predictive models to address spatial health risks, showcases the wide-ranging applicability of predictive analytics across various sectors. In addition, by analyzing the impacts of AI in both healthcare and geospatial intelligence, we not only advance technological capabilities but also enhance the health and safety of our population.

School of Engineering and Applied Science

Bachelor of Science in Computer Science

Technical Advisor: Briana Morrison

STS Advisor: Joshua Earle

Technical Team Members: Jane Chu