j# Smart Salinity Classifier + Gemini AI

Water quality meets explainable AI—an intelligent way to predict, understand, and act.

Hello to the Dev team!!!!!!!

Inspiration I got the inspiration from the situation in Delhi India and a drive to help researchers,farmers and ecologists

2.What it does This app is here to simplify your data and we use Gemini AI with Salinity Classifier where you add a csv file and bam you get the roc curve,confusion matrix which can explain the patterns to ecologists pro and tech noobs 3.How I built it We built it using Streamlit for UI Google Cloud for app launching Rocket+Ridger Classifier for Data Simplification Matplot.lib for plotting ROC curve and Confusion Matrix Logistic Regression buit to output probabilities for classes

4.Challenges I ran into As a solo fresher dev in my first year in college enjoying my summer vacations with a i3 11th gen I can say Proudly that after all these sleepless nights I am proud that i made this app

5.Accomplishments that I am proud of

Fully solo-built project from logic to deployment
Smooth integration of two powerful AI systems: ML + Gemini
Gemini's ability to explain each parameter live in context
Clean, interactive design with real-time predictions and explainability

5.What I learned

How to combine LLMs and classic ML for decision intelligence
Prompt engineering techniques that enhance LLM explanations
Streamlit’s deployment flexibility and GCP integration workflows
The importance of interpretability in AI tools—not just accuracy Trained on Thorslund.el(2021)dataset

6.What's next