Hi, I’m Ishita.
Data analyst. Storyteller. I like talking to people.
Also: tea drinker, code wrangler, and positive thinker.
After finishing my PhD in Cancer Biology, I decided to improve my German and thus attending courses currently for B2.2 level at vhs. While attending language classes, I’m also diving into new skills and having fun with personal projects (because why pick just one challenge?). I’m curious, adaptable, and always learning. This portfolio is where I share my work, my interests, and the projects that keep me excited about what’s next.
Thanks for stopping by!
My Projects
Confess2PopeAI
A fun project building AI chat box with gemini-2.0-flash
You can now confess your sins anonymously to the AI Pope. He listens without judgment, shares relevant Bible verses, and offers spiritual guidance to help you move forward in life. This project blends playfulness and purpose — showcasing the potential of generative AI for empathy, reflection, and healing.
Tech Stack & Tools Used
- Python – Backend logic & API proxy with Flask
- Vertex AI Studio (Google Cloud) – Custom-tuned prompt management with Gemini
- HTML, CSS, JavaScript – Frontend integration
- Postman – API testing
- Git & Bash – Version control & local development
Credit Score Calculator
This project calculates credit scores using Python and stores them in a MySQL database.
A Python project that simulates credit scoring using behavior-based metrics and stores the results in a MySQL database. The scoring logic is inspired by FICO-style weighting, using a simple credit score formula. The calculated raw score is then scaled to a range of 300–850 to mimic real-world credit scores. The project features automatic schema updates (adding columns if missing), weighted metric processing using Pandas, and full SQL integration using MySQL Connector.
Tech Stack & Tools Used
- Python – Data processing with Pandas
- MySQL – Database storage
- MySQL Connector – Python-MySQL integration
- Git & GitHub – Version control and code hosting
Predicting Wine Quality with ML
A Python-based project that uses the UCI Wine Quality dataset to compare two classification models: Decision Tree and Random Forest. Based on a hypothesis that Random Forest would offer better generalization and accuracy, both models were trained and evaluated using accuracy as the primary metric. The project also includes a feature importance analysis to determine which physicochemical variables most influence wine quality. Results are visualized using Pandas and Matplotlib.
- Tech Stack & Tools Used
- Python – Data analysis and model training using scikit-learn and Pandas
- Jupyter Notebook – Interactive code and results presentation
- Matplotlib – Feature importance visualization
- UCI Dataset – Real-world wine quality data obtained from kaggle
Blog posts
Testimonials
“Frau Dr. Parui ist in hohem Maße teamfähig. Sie hat eine sehr angenehme Art der Interaktion und eine große Bereitschaft, sich für die Interessen der Laborgemeinschaft einzusetzen. Sie war an der Betreuung naturwissenschaftlicher und medizinischer Doktoranden und Bachelorstudenten beteiligt. Ihr Verhalten gegenüber Vorgesetzten, Kollegen und externen Kooperationspartnern war stets einwandfrei.”
– Principal Investigator Prof. Dr. Georg Häcker, Uniklinik Freiburg