Computer Science
Mrs. Taricco

About The Class

Computer Science at MAMS is taught by Mrs. Taricco. For A Term, we worked on HTML and CSS skills that were used to develop websites, such as this one. Since I took AP Computer Science coming into MAMS, I worked on the advanced sequence and completed advanced programs. These programs looked to solve problems relating to mathematical concepts, such as the Fibonacci sequence, prime numbers, and number theory.

Lyric Generator

During B Term, Svabhu Govindaraj, another advanced student, and I worked on an independent CS project, where we were able to explore different branches of CS and decide what we wanted to explore for our project. We decided to create a machine learning algorithm that could generate lyrics that follow the flow of the artist that the user picked. To do this, we coded an LSTM machine learning algorithm on Jupyter Notebook that looked at the past three words in the lyrics to generate the next one. The program was trained for 30 epochs on training data that consisted of lyrical data of the artist. From here, the program was integrated with a GUI application window that was coded using Anaconda Spyder. Here are some generated lyrics for Taylor Swift that were created by the model. Below is the Python code for the machine learning model.

Apps For Good

During D Term, we developed Android Applications using Android Studio. I worked in a group with Alexis Chong and Isaiah Bateman to develop the app, PETential.

Target Audience

This app looked to solve the problem of the increasing amount of surrendered pets and pets in homeless shelters. Due to the pandemic, many people adopted pets due to impulsive decisions without first considering if they would be able to provide the needed items for a pet. Our app looks to combat this by matching people with pets based on purely quantative information such as their budget and their home square footage to ensure that the pets they are recommended are viable based on numerics rather than any subjective opinions.

MVP

For our minimum viable product, we wanted our app to allow users to input their personal data to get a pet recommendation that is catered to their information, save user information in an editable profile, and provide information on a place near them where they can adopt/buy the recommended pet.

Final Product

To create this app, we first did research on pet data in the United States and homeowner statistics for different pets. This data was then used to create an algorithm that would use purely quantative data to match a user to a pet. This algorithm was implemented through a weighting system that gave certain pets points based on each user's information. The layout if our app is shown in the poster below and was designed in Android Studio. We worked for 2 months with Android Studio for the final set up of the app, including the UI and adding a Google Maps implementation to find pets near us.