Side quests in code
Soft Projects
AREAS OF CURIOSITY
These are my soft projects – part code, part sketchbook. Most of them were born out of class prompts, personal curiosity, or simply the desire to try something new. They may not be polished products, but each one taught me something about logic, structure, and digital storytelling. Some were experiments. One became an app I’m proud of. Others are here simply because I had fun building them.
01.
P5JS Experiments
During my Creative Coding class, I spent a lot of time with p5.js – a JavaScript library that turns code into canvas. These projects were my way of exploring logic, visuals, and interaction through simple sketches. Some loop, some react, some exist for the sake of trying something new. None of them is perfect, but each one helped me see code as a creative tool, not just a technical one.
02.
My Sticker app
What started as a birthday present turned into one of my favorite personal projects. I designed and built a simple sticker app – a playful space to collect and share my hand-drawn illustrations. It was a mix of creativity and code: drawing each sticker by hand, organizing them into themed sets, and designing a clean interface to bring it all together. This app wasn’t built for the app store – it was built for joy.
03.
YouTube Transcript Summarizer
An algorithm to summarize YouTube videos efficiently.
In a world flooded with videos, time often limits our ability to consume all the content. Hence, it’s crucial to summarize information for easier digestion without losing its essence. My goal is to create an algorithm that can efficiently summarize YouTube videos by extracting key text from their transcripts and refining it with a thesaurus. This automated approach lets us swiftly identify crucial patterns in videos, saving time and effort otherwise spent on combing through the entire content.
04.
Detecting Fake Job Postings
Predicts job outcomes accurately using advanced machine learning techniques.
Our mission is to achieve the utmost accuracy in predicting job outcomes-whether genuine or fake-by employing cutting-edge machine learning techniques. Our approach involves applying Supervised Machine Learning Techniques (SMLT) to comprehensively analyze the dataset, including variable identification, handling missing values, and data validation. We’ll ensure the dataset is meticulously cleaned and prepared, accompanied by insightful data visualizations. The finale involves crafting an ensemble model utilizing ML Algorithms such as XGBoost, SVM, Logistic Regression, and Random Forest Classifier, leveraging the top 4 contributing features. To showcase the efficacy, the model will be integrated into a Flask application for a live demonstration.
Team Size: 3
Team: Aadharsh K Praveen, Harsita R, Rachanna Deva Murali
Professor: Niveditha S
Paper Published
