Halma Strategy Board Game with AI

• Python
• Pygame Library
• Git, Github

This is a Python implementation of the board game Halma, using the Pygame library for the graphical user interface. Halma is a two-player strategy board game invented in the 19th century. This version of the game is played on an 8x8 square board between a player and the computer. Similar to the game Chinese Checkers, gameplay involves players taking turns moving their pieces from their starting corner of the board into the opposite corner before their opponent.

I decided to build this project after finishing my second quarter as a CS student at Oregon State University. My motivation for the project was to apply some of the heavy concepts I've learned so far in my coursework towards a tangible working product. Working on this game pushed me to get my hands dirty with larger scale Object-Oriented Programming, recursion, modules, and thorough documentation as well as the opportunity to learn more about AI.

While I ran into many challenges while working on this project, the two biggest ones were move validation and evaluating states of the game board to enable proper decision-making by the AI.

• Enforcing proper move validation was the biggest hurdle I had to overcome during this build. In Halma, players may move a piece to any empty adjacent square in any of the eight directions surrounding their piece. If such a move is made, their turn ends. Players may also make a jump over any adjacent piece of any color into an empty space on the other side of that piece. Multiple jumps can be chained together in a single move as long as each jump is over a different piece and into an unvisited space. The solution came down to building a list of all valid moves into adjacent spaces with a helper function and adding all valid jumps to that list using a separate recursive function.


• The other main challenge I faced was writing an evaluation function that properly informed the AI's movements based on the current state of the game board. While I tried out different factors such as central control of the board, and blocking opposing pieces, I eventually settled on evaluation based on three factors:
  
  
  1. Distance between each player's pieces and the opposing starting zone
  2. Number of pieces in the opponent's starting zone
  3. Penalty for pieces remaining in their own starting zone
  
  The three factors are determined as numeric values and totaled as a floating-point number representing the evaluation score. The final stage of development for the project as a whole came down to playing test games against the AI and adjusting the weight of these three factors to facilitate the strongest decision-making by the AI.

Vocabulary Worksheet Generator: Spring 2023 Hackathon Project

• Python
• Flask
• Bootstrap
• OpenAI API
• Heroku

This was a project I did as a team in the Spring 2023 Beaverhacks Hackaton at Oregon State University. The theme of the hackathon was leveraging AI. My team and I decided to build a tool for English teachers, like myself, to create a tool that could provide real-world value. Utilizing the OpenAI API, our app generates vocabulary worksheets for teachers based on unique user input.

The app features an explanatory index page and a button to take users to a form. The form page allows users to provide input on the type of worksheet they want to generate based on:

• Grade level
• English ability
• Unique vocabulary
• Question type

Once the user completes and submits the form, a call is made to the OpenAI API, passing the form input as part of a prompt, and a worksheet is generated in a visually pleasing format. The user is routed to a download page where they may preview their .PDF worksheet and download or print it.

While our team ran into many challenges working on the project together, such as working collaboratively in Git and working across multiple time zones, I will highlight some of the specific challenges I faced in my own share of the work:

• As this hackathon was during my third quarter as a CS student at OSU, I had worked purely in Python in my coursework up until this point. Learning Flask syntax, templating with Jinja, and basic HTML and Bootstrap CSS was a lot to take on for a two-day project. It was a lot of time spent digging through documentation, but it provided a valuable opportunity to learn new skills and frameworks and actually build something tangible and useful.


• The other main challenge I faced was working with my team to figure out how to take the API response from OpenAI and present it to the user as a .PDF file to be downloaded or printed. A bit of searching led us to try out a Python library called FPDF which allowed us to set up a customized .PDF populated with the output from OpenAI.

Upon finishing our app, my team and I successfully deployed it on Heroku, where we had it live for the duration of the hackathon judging, but took it down shortly after, so as not to rack up expenses. The project was a valuable way to build something as part of a team, learn new technologies, begin to understand server and router functionality and API's, and deploy a project for the first time.

Flask web application with Bootstrap CSS

• Python
• Flask
• HTML
• Bootstrap CSS
• Git, Github
• AWS Elastic Beanstalk

I created this website to serve as a personal portfolio for my professional career as a software engineer after completing my first year as a CS student at Oregon State University. Up to this point, my only web development experience was an Intro to Web Dev class focused on the MERN stack. With this app, I wanted to challenge myself to practice the concepts I learned from that class using different technologies.

Through creating this project, I am happy to take away a much stronger grasp of creating applications with Flask and frontend styling with Bootstrap CSS. I had minor experience using those technologies in a previous hackathon project, and wanted to get my hands dirty with them again and really understand them better. It was a great exercise in reading documentation and acquiring new skills in order to create the portfolio website I envisioned.

While developing this project, I encountered several significant challenges that provided valuable learning opportunities. The primary focus was on setting up an efficient deployment environment on AWS Elastic Beanstalk and establishing a seamless CI/CD CodePipeline from GitHub.

• One of the main challenges I faced was configuring the deployment environment on AWS Elastic Beanstalk. This involved creating a new user and role manually, properly configuring environment properties ensuring the security of sensitive information, and optimizing the environment's performance to handle the website's traffic effectively.


• Additionally, establishing the CI/CD CodePipeline demanded a thorough understanding of integrating GitHub with AWS CodePipeline. I had to ensure that each code commit triggered automatic deployments, enabling continuous integration and delivery while maintaining version control and preventing errors.

Overcoming these challenges has not only enabled me to successfully deploy my portfolio website but also honed my skills in AWS, CI/CD pipelines, and infrastructure management. I am now better equipped to tackle future projects with confidence and deliver higher quality results.

Jupyter Notebook Data Analysis: Winter 2023 Hackathon Project

• Python
• Jupyter Labs
• Pandas
• Numpy
• Seaborn
• Matplotlib

I worked on this project as part of a team for a Health and Healthcare Hackathon, where we sought to address mental health and wellness, an increasingly prominent topic in health discussions. My team aimed to contribute value and problem-solving to the healthcare field by using data analysis to understand how music connects to mental health. Through our data analysis, we aimed to identify potential correlations between an individual's music tastes, listening habits, and self-reported mental health. The goal was to provide valuable insights for professional music therapists, enabling them to better comprehend the relationship between music and their patients.

We initiated our data analysis process by acquiring a relevant public dataset from Kaggle and cleaning and analyzing it using Python in Jupyter Labs. Employing various Python libraries such as Pandas, Numpy, Seaborn, and Matplotlib, we skillfully manipulated the data and crafted insightful visualizations. Our findings were thoughtfully summarized, providing valuable insights that can contribute to the field of music therapy.

Participating in a hackathon can be stressful and challenging as you are part of a collaborative effort to build something in a short time. My team ran into a number of challenges during our development process:

• Setting up the coding environment, particularly Git and GitHub, posed several challenges for our team.
• Initial plans to create an interactive dashboard with widgets using Panel faced difficulties, leading us to focus on static charts created with Matplotlib and Seaborn.
• Accessing and isolating specific data from the CSV file for charting was a hurdle.
• The dataset's limitations in terms of varied numerical data constrained our data exploration possibilities. Additional demographic columns and a larger survey pool would have provided more insights.

Overall, I am proud of my team's ability to come together to develop this project on the heels of finishing our first ever Intro to Computer Science course together at OSU. Jupyter labs was new to us all, we knew very little about Git and Github, and we had little to no experience using additional python libraries such as Pandas, Seaborn, and Matplotlib. We achieved working experience reading through documentation for new technologies and contributing to a collaborative project in a time sensitive environment.