Projects

SPX500 Class Action Stock Trends

I worked as a frontend engineer on a team to develop a dynamic React-Flask website to analyze SPX500 companies' stock trends following class action lawsuits. I designed the frontend using Node.js and Bootstrap, implementing features such as search, sort, and filter. We built a REST API with Postman and populated it with SPX500 stock and legal history pulled from relevant APIs. This was used in turn to populate our GCP PSQL database. Interaction between our database and the flask app was facilitated via SQLAlchemy. A demo video of the website is shown below.

COVID-19 Tracking App

I developed an app using Kotlin and Gradle in Android Studio to track traces of COVID-19 within a certain city or state in the United States. Enriched recycler views were used to allow for fluid and seamless UI navigation in the app. One fragment in the app included a heat map of COVID-19 cases within a selectable state. This heat map also included the implementation of long click listeners within a certain tolerance of markers on each city to generate a toast or preview of certain statistics within the county. Another feature of this app was the construction of a GPS system using the Google Maps API so that the user's current location could be detected automatically. A demo video of the app on an android emulator is shown below.

Pharmaceutical Data Visualization

After my insurance provider substituted its coverage of my prescription for Descovy with its generic counterpart Truvada, I worked with a friend to construct multiple data visualizations to compare the risks and costs of the two prescriptions. I used Pandas to extract data from several csv files pulled from various academic journals, and I used Altair to craft the visualizations from the data. These visualizations include a vertically juxtaposed stacked bar chart with a selection tooltip to compare the rates of side effects for the two drugs. I also implemented a choropleth with an Albers Projections to compare the manufacturing costs to revenue ratio for the two drugs per county in order to determine whether the far higher price of Descovy was justifiable. These visualizations are shown below. Some of the data can be hovered over or selected for additional information. More information regarding the sources of our data, the classification of the data, and the methods used in creating these visualizations can be found in this report.

Seizure Prediction Analytics System

I worked on a bioengineering team in Texas Engineering World Health at UT Austin to develop the software components for an electrode headset to warn seizure patients prior to seizure onset. 30 seizure patients agreed to have their neural oscillation readings used for training our algorithms. This data was manipulated with Python and Pandas to fit the Deep Convolutional Auto-Encoder and Long Short-Term Memory (LSTM) neural network systems. The auto-encoder was used to reduce noise and outliers from the data and the output was sent to the LSTM to develop an artificial "memory" which would allow the algorithm to recognize patterns and grow more accurate when trained with more data. This project was submitted to the international Engineering World Health Competition and achieved 3rd place. The problem definition and statement of impact can be viewed here.

Projects with Underserved Communities in Siripudi, India

I worked on a team of 7 college students to fund and design a community center for a small village in India. We raised $30,000 and fully designed the community center for the implementation by our contractor in May of 2022. Some photos of the building can be viewed here, and our fundraising video can be viewed here.

Some of these projects and some smaller brain teasers can also be viewed on my GitHub profile or my resume where my experience and education are also recorded.