Small animation of me coding

Computer Science

Our computer science class is taught by Mrs. Taricco. This class focuses on fundamental concepts of object-oriented programming and methodologies with a focus on problem solving skills. Web development with CSS and HTML are taught, as well as Java and app development. Below are some examples of my work.

Apps for Good

During D Term CS, I worked with Anne Tie and Nihitha Reddy to develop an app titled LaunchGuide. We worked with Jack Peacock, a partner of the WalkFit program designed to help increase awareness surrounding visually impaired people (VIPs). We specifically identified an issue with crosswalk signals.


Individuals who are blind or suffer from low vision, hereinafter referred to as VIPs, are usually fearful of venturing out on their own due to the environmental obstacles present outdoors. When VIPs do venture out, they often encounter many obstacles, especially when crossing the street. When crossing intersections, often they have crosswalk signals that display a white walking man or a red hand to indicate whether it is safe to cross. However, these signals create challenges for VIPs as they often lack audible information that tells VIPs their status with respect to crossing the street. Since it is imperative to improve accessibility, our project aimed to create an app that assists VIPs when crossing the road by providing auditory signals.

Apps For Good Poster

Problem and Target Audience

Globally, at least 2.2 billion people have a near or distant vision impairment (Why Accessibility Is Important | National Center on Deaf-Blindness, n.d.). Those who are visually impaired experience partial or complete loss of vision that causes problems not fixable by conventional devices, such as glasses (Blind vs. Visually Impaired, 2019). Today, as intersections become more and more congested, signaling schemes often follow suit and become more complex. In the past, the standard design parameters were based on an able-bodied person, or one with good vision, hearing, and mobility (Accessible Pedestrian Signals: Understanding How Blind Pedestrians Cross at Signalized Intersections, n.d.). Unfortunately, these systems often fail to provide sufficient non-visual information for crossing decisions by VIPs. This is due to the information barriers that restrict an individual’s ability to recognize and receive information from their surroundings. More importantly, it prohibits them from using this information to decide on a course of action. As 26 percent of all Americans have some type of disability, and with that percentage increase, it is imperative to require design parameters to meet the needs of all pedestrians (CDC, 2023).
VIPs can travel and cross streets with human guides, white canes, guide dogs, and many other methods. Regardless of the aid that is used, street crossing often includes many steps. First, the pedestrian must determine if they have reached a street. Often, they use a combination of cues, such as the curb or slope of the ramp, traffic sounds, and detectable warnings. Next, they must recognize the exact street they have arrived at. This information is not typically provided in an accessible format, so VIPs must develop a mental map or seek assistance from other pedestrians. If the VIP has identified that they have come to an intersection, they must obtain critical information about intersection geometry, which is comprised of the location of the crosswalk, the direction of the opposite corner, the number of intersecting streets, the width of the street to be crossed, and whether there are any islands or medians in the crosswalk. In order to determine these details, they listen to vehicular sounds, traffic patterns, and search the sidewalk area for poles with pushbuttons (Can I Cross the Street? Considerations for a Blind Pedestrian | NADTC, n.d.). Unfortunately, it has become difficult to determine the type of traffic control at intersections that may fail to access the pedestrian push button and crossing at times other than the pedestrian phase. After determining the layout of the intersection, aligning to face toward the destination curb, determining that the intersection is signalized, and having pushed a button (if available), VIPs must recognize the duration of the walk interval. Ultimately, VIPs have many problems at these intersections. Since many only provide limited auditory signals, it can be difficult for a VIP to determine whether it is safe to cross, how much time they have to cross, or where the other side of the road even is.
It is imperative for ubiquitous appliances, such as pedestrian crosswalks, to be accessible. Accessibility is the concept of whether a product or service can be used by everyone however they encounter it (Why Accessibility Is Important | National Center on Deaf-Blindness, n.d.). Not only does accessibility provide vital parts of user experience design, it also often benefits all users. These crucial statements have also been reiterated through the Americans with Disabilities Act of 1990. Specifically, the law outlines that the WALK/DON’T WALK cues from the visual pedestrian signal heads should be conveyed to pedestrians who don’t have visual cues (Barlow, 2009).

With this in mind, we developed an app that would be able to aid VIPs when crossing intersections that have visual signals.

Our App

Working with Mr. Peacock, we quickly cemented the issue that when VIP’s try to cross the street, they often lack enough details to feel confident crossing alone. Because of this, we developed an app that scans the user’s surroundings in real time. It utilizes a light-weight on-device binary image classification model that is able to scan and identify crosswalk signals such as the white person and the red hand. If nothing is detected, the app doesn’t return anything. However, if a signal is detected, the app will alert the user. If a walk signal is detected, it will return a “ding” sound, letting the user know it is safe to cross. However, if a don’t walk signal is detected, the app will return a “wait” sound to let the user know they should wait.
Using Android Studio, we first implemented the camera function using the CameraX API. This API accessed the user’s camera to provide the app with a stream of images to be analyzed. Specifically, we used the CameraX Image Analyzer. From there, the images are passed into our machine learning model that was programmed in Python with Sci-Kit Learn and Keras, and then converted into a Tensor Flow Lite model for use with Android Studio. If the model returns a confidence of 80% or above, the app will alert the user. If a walk signal is detected, the main screen will prompt the user to walk, as well as provide an auditory alert. The same is done for when a do not cross signal is detected.
This was our minimum viable product. However, we also implemented other pages and features for user quality of life, since we particularly wanted to emphasize accessibility. Firstly, our app had two themes: a dark theme and a light theme that followed the user’s system preferences. We made sure to use high-contrast and VIP-friendly color palettes to allow users with limited vision to still navigate through the app. In addition, we implemented an intuitive and user-friendly way to ask for permissions, ensuring that at no point would a user solely relying on audio get lost. Furthermore, our app is fully compatible with TalkBack, the Android feature that allows for text on screen to be read aloud. Upon opening with TalkBack, all text is automatically read, and buttons have appropriate and intuitive descriptions. Finally, we implemented help screens that tell the user exactly what to expect from the app.

Lab 06 - Line Art

While learning Java, we had labs which were assignments where we had to use our knowledge and skills to create a program to perform a specific action and show proficiency. This ranged from calculating taxes based on different incomes, to drawing an illusion. This lab was to allow us to show our knowledge of loop control structures and the Graphics class. For the line art lab, we were tasked to draw straight lines in a rectangle connecting two adjacent sides. These lines were to be drawn so that the starting points and ending points were equidistantly spaced along their respective sides. Though each line was a straight line, the end result looked like a curve, since there were so many lines. For even more of a challenge, we also had an option to draw a smaller version of the rectangle inside our first rectangle. Since so many lines had to be drawn, the process had to be automated using a loop control structure, such as a for loop. This loop essentially keeps repeating a task until a certain condition is met. In this case, the task was to draw a straight line, and then move a set distance, and then repeat until it reached the end of the rectangle. In addition, I added extra code that slightly changed the color each time. Here is the result. I especially love how the colors turned out.

image of lineart lab


Aside from labs, we also did many smaller exercises to use our knowledge to code various tasks. One such task was to code a program that would draw ten different stars of random sizes in random places. This was done using a combination of algebra, trigonometry, and geometry, as well as static arrays in Java. A static array is essentially a list that holds a sequence of values. In this case, this was the x and y values of each vertex of the star to draw the shape using a polyline. First, a random center point for the star was generated, and then a random radius. From here, the locations of the rest of the vertices were calculated, and the star was drawn. This was repeated ten times. The end result changes every time the program is run, especially because I decided to make each star a random color as well. However, this shows what running the program might look like.

image of stars assignment