In Computer Science, we are learning the foundations of different coding languages. It is important to understand Computer Science in a world where AI is an unknown quantity which could potentially revolutionize the way we live as it develops. As of right now, I am coding this very line in HTML for my CS Website project.
This was the second lab we did in computer science class. The task was to convert time given in milliseconds to time in hours, minutes, seconds, and milliseconds.
This lab emphasized algorithmic thinking and step-by-step problem decomposition. I learned how to use integer division and modulo operations to break a large value into meaningful components. The assignment reinforced precision in variable handling and output formatting. It also highlighted how even simple problems require careful attention to logic and edge cases.
As part of the advanced cs curriculum, I must work on an independent year long coding project. This is my MVP, or the a working yet not finished version of my project. Scroll down to the README section to learn more about my project!
Developing the MVP required me to prioritize core functionality over additional features. I focused on building a stable foundation that could be expanded through iteration. This process taught me version control habits, debugging strategies, and time management. The project reflects my ability to independently plan, implement, and refine a long-term software idea.
Apps for Good is a team project where small groups design and build a mobile app that solves a real-world problem. My group consisted of Ananth Gomattam, Parnitha Karapakula, and me, with Mrs. Taricco as our advisor. We built ForSeizure, a predictive epilepsy risk management app whose tagline captures the idea well: helping you foresee seizures.
Epilepsy affects 1 in 26 people globally, but the hardest part of the condition is unpredictability. Not knowing when a seizure will happen forces patients to limit driving, travel, exercise, and many other daily activities. The problem is even greater for the roughly 1 in 3 epilepsy patients with drug-resistant epilepsy, where no medication combination stops the seizures. For those patients, the risk of Sudden Unexpected Death in Epilepsy (SUDEP) is significantly higher.
Existing apps like Epsy, Seizure Tracker, and EpiDiary mostly work as retrospective logs. They record what happened after a seizure and organize that data for doctors, but they do not help patients prepare for the day ahead. None of them combine lifestyle factors with environmental data to forecast risk before it happens. ForSeizure was built to fill that gap, and our team was motivated in part by a teammate's relative who lives with epilepsy and unknown triggers.
The primary audience is young people with epilepsy between the ages of 13 and 30. They are comfortable with mobile health tools and feel seizure unpredictability most directly because it affects school, work, social life, and independence. A secondary audience includes caregivers, parents, and clinicians who track safety or evaluate treatment decisions. The app is designed so all three groups can use the same data without separate workflows.
ForSeizure moves epilepsy management from retrospective record-keeping to proactive risk navigation. The app collects two kinds of data: daily lifestyle inputs such as sleep, stress, diet, drug use, and hormonal changes, plus environmental data like barometric pressure, temperature, and air quality. A weighted mathematical model combines these inputs with the user's seizure history to generate a daily safety score from 0 to 1. A score of 0.35, for example, means the model considers it a high-risk day and sends a notification so the user can plan accordingly.
Our MVP focuses on trigger identification, which the prediction model depends on. It delivers three working screens: a home dashboard with today's risk score and recent logs, a daily entry form for seizures and lifestyle factors, and a medication log for prescriptions and timing. The prediction algorithm and safety score already work in the MVP, and they improve as the user adds more data over time.
The core of ForSeizure is a five-step prediction algorithm. First, the app labels each logged day as a seizure day or a normal day. Second, it runs a trigger analysis: for users with fewer than ten seizure days on record, it uses simple threshold comparisons; with more data, it runs Welch's t-tests and ranks confirmed triggers by effect size. Third, the app runs an ANOVA to detect interaction effects, because some triggers are more dangerous in combination than individually. Fourth, it calculates the day's risk score by summing the weighted deviations from the user's personal baseline. Fifth, all weights and baselines are recalculated whenever new logs are added, so the model stays personalized.
For storage, ForSeizure uses SQLite locally for recent logs and offline entry, with Firebase as a secondary layer for long-term history and cross-device access. All stored data is encrypted, and users can delete any entry at any time. Compared to competitors, the value proposition is simple: Epsy and EpiDiary show users what happened, while ForSeizure shows users what may happen.