Unified ETL Pipeline for patients statistics dashboard
Built a real-time optimized ETL pipeline for distributed systems reducing data loading time by 15 minutes and an adaptive patient forecasting system
Technologies:PostgreSQL, Kafka, PySpark and Power BI
Automated Invoice Processing and Predictive Inventory Management
Implemented YOLO V8 and OCR technology to automate invoice data entry, reducing manual processing time by 60 hours per week and improving data accuracy by 85%.
Technologies: YOLO V8, OCR, Python, TensorFlow
Brain Tumor Detection Using Deep Learning with XAI
Developed a CNN-based deep learning model to classify MRI brain scans into glioma, meningioma, and pituitary tumors. Achieved 96.5% accuracy using an cnn model
Technologies: Python, TensorFlow, Keras,LIME, Grad-CAM
Business Dashboard
IDesigned an interactive Power BI dashboard to monitor sales trends and analyze customer behavior for a metal manufacturing business. Enabled dynamic filtering by customer, date, and product category to uncover key patterns, top-performing clients, and peak sales periods.
Technologies:Power BI
ABC Gold Procurement Strategy Dashboard
Designed a data-driven procurement system using GARCH, Bollinger Bands, LOF, and K-Means to analyze gold price trends and customer behavior. Visualized insights with Power BI.
Technologies:Python, Scikit-learn, Pandas, Statsmodels, Power BI, Plotly
Earthquake Analysis Dashboard
Built an interactive Streamlit dashboard to visualize global earthquake data (1995–2023) with filters, geospatial heatmaps, and statistical insights. Integrated anomaly detection and trend analysis.
Technologies: Python, Streamlit, Pandas, Plotly, Folium
Agricultural Intelligence Web App
Built a Flask app that recommends crops and fertilizers from soil data and detects crop diseases using machine learning and deep learning models.
Technologies: Flask, Python, Scikit-learn, TensorFlow, Keras, HTML/CSS
Modules:Crop Recommendation, Fertilizer Suggestion, Disease Detection
Movie & Book Recommendation Using Twitter Sentiment Analysis
Built a personalized recommendation system using Twitter sentiment analysis and Multinomial Naive Bayes to suggest movies and books based on user tweets. Integrated with Streamlit for live interaction.
Technologies: Python, Streamlit, Tweepy, Scikit-learn, NLP
Algorithms Used: Multinomial Naive Bayes, Genre-Sentiment Mapping
Secured departmental store management
A system for efficiently managing supermarket operations using computerized tools to streamline processes, store data, and enhance staff productivity.
Technologies: c++
