PaidinToken
My PaidinToken project was my first full-stack project. I’ve always been interested in cryptocurrency and hip-hop music. So when Artist MoneyMan was paid a $1 million advance in November 2021 in Bitcoin instead of fiat dollars, I searched online for a site that would display the real-time value of the large payment based on the real-time price of Bitcoin. I committed to creating it once I saw that this didn’t exist. I researched ten large payments made to celebrities and used the investment amount and time of investment to calculate the price paid per Bitcoin and the number of Bitcoins. I formed the table and display in HTML and CSS and constructed the updating numbers as the new price of Bitcoin came in every minute in JavaScript. For the backend, I utilized PythonAnywhere, where I wrote a Python script that would use an API to pull the price of Bitcoin, update a CSV file, and send it to my AWS S3 Bucket every minute. My AWS S3, Route53, and CloudFront configurations allowed PaidinToken to be hosted and constantly updated for its 1,000 unique pageviews in the first month.
It was my first major project that I took on by myself because of how interested I was in the unique investment made by an artist I listened to. When I realized that other celebrities are accepting large payments in Bitcoin, I knew I wanted to share this knowledge with others. I have been blessed with a family knowledgeable about personal finance and different means of having one’s own money and making more money for them. I wanted to share this by allowing people to see celebrities they may have a personal connection to making investments in unique assets such as Bitcoin and seeing how profitable they were in hopes of allowing more people to consider diversifying their investments. Ironically, I pursued this project, and it became public at a time when the price of Bitcoin decreased. Nonetheless, I am grateful for the experience as it taught me numerous lessons, technologically and personally.
In my experience developing PaidinToken, I’ve picked up plenty of lessons. Initially, I iterated through numerous services, backends, frameworks, and versions before settling on how PaidInToken is displayed now. In my pursuit of a means of running automated workflows, I experimented with services like GitHub Actions before deciding on Python Anywhere. I circled through different CSS frameworks like Bootstrap before choosing Bulma. In this experience, I also spent much time learning how to choose and pursue a tech stack I would work with. In this, I spent a significant portion of time learning how to learn from online sources. I worked on this before ChatGPT, so I had to find people who had encountered similar issues and try to resolve them with what I saw worked for them. I was fortunate to have built a strong relationship with a mentor, Rob, at Stacker, who offered me advice and direction. With all of these varying resources, I was able to launch in August of 2022.
Last Next
Automated Testing Software
I’ve worked on and with Automation Testing Software daily in my internship with Stacker Media. Our data-driven story-creation process relied heavily on the ability to access and collect data from online sources. So the automated approach would be to utilize Python’s Request library but this limited what sources we could use because some pieces of data exist in a URL that can only be accessed after a string of actions. For example, as a sophomore, I wrote a script to follow and unfollow people on Instagram, but there’s no 1 URL I can load to find this. I would need to imitate the human interaction with the computer and this process. The Selenium Framework allows me to do just this. I’ve created varying sorts of bots such as sneaker bots which are highly desired by many who are consistently unable to checkout fast enough for highly sought-after and lucrative shoes or clothing. The bot would intake a person’s desired item, size, and personal details to then imitate the process of finding the item, adding it to one’s cart, and checking out in a way that supersedes a casual person’s speed. I refined it to make it to overcome bot-stopping CAPTCHAs by temporarily pausing the bot to prompt the user to complete it themselves to which the bot would immediately complete its process. The accompanied video shows this.
Though projects like this were for my entertainment and spawned from my interests, the bulk of my experience with automated testing software came from my internship experience. Writing countless scripts that would run for minutes and sometimes hours on my MacBook Air pulled heavily on my CPU, impaired my ability to simultaneously complete other tasks while running a long script, and didn’t run to the best of its ability. In researching ways to improve this, I learned how to run automated scripts in a headless browser which reduced CPU usage, allowed me to use my computer while running scripts, and ran faster. This empowered me to collect thousands of pieces of data in less time, thus producing more data heavy stories without reducing my computer’s speed. I also refined my skillset with automation testing software by learning how to navigate the new content I would land on. Plenty of times have I landed on a URL where I’ve collected its HTML where an extensive portion of it was JSON code that I learned to parse to get to the desired data
All in all, I’ve worked with Selenium through a variety of scopes. In large part, I’ve collected data that I would then clean and convert into a new valuable product for customers in my internship experiences. The data I’ve collected has also served my sports analytics projects and class assignments. Selenium has also automated processes for myself and my work. Having used the framework to resolve various technical tasks, it is one of my strongest skills.
Last Next
Bitcoin Analysis Project
This is a collaborative project alongside a classmate, where we took the data from PaidinToken to analyze the price of Bitcoin from various perspectives. The way I set up PaidinToken, I had 1 file constantly updated every minute with the current, day-high, and day-low prices. The file was continually the same size because those three numbers were overwritten and updated. For this project, we wanted to analyze an extensive portion of Bitcoin data. So I changed the structure of the backend to instead add onto the previous data. So we ended up with over 56,000 rows of data for 11 rows of data. We took this data and analyzed it through the lens of the highest and lowest prices, the time trend for these prices, outliers, average changes, correlation analysis, and real-world factors. Please feel free to read more about this in the project’s official PDF site
Last Next