{ "meta": { "source": "prasadgade.dev", "generated": "2026-04-19", "version": "1.0", "description": "All projects by Prasad Gade with descriptions, tech stacks, and links" }, "data": { "projects": [ { "title": "IPL Analytics Platform", "subtitle": "Interactive Cricket Data Dashboard", "category": "Data", "stats": "18 Seasons", "description": "3-stage pipeline over 278K ball-by-ball records with 14 dashboard pages, custom SQL query builder, and validated IPL aggregates.", "detailed_description": [ "Engineered a 3-stage data pipeline processing 278,205 ball-by-ball records across 18 IPL seasons (2008-2025), producing 90 derived features and 15 aggregate parquet files", "Built 14 interactive dashboard pages with 58 preset SQL queries, a custom query builder supporting 9 entity types", "Achieved zero data mismatches across 703 batters and 550 bowlers; validated with 25 automated tests" ], "tech": ["Python", "Streamlit", "DuckDB", "Plotly", "Pandas", "NumPy", "PyArrow", "Apache Parquet", "SQL", "pytest"], "github": "https://github.com/prasad-gade05/IPL_Analysis", "demo": "https://analytics-ipl.streamlit.app/", "dataset": "https://huggingface.co/datasets/prasad-gade05/ipl-enriched-2008-2025", "kaggle_dataset": "https://www.kaggle.com/datasets/prasadgade/ipl-2008-2025-enriched-dataset" }, { "title": "UPI Analytics Platform", "subtitle": "Data Engineering & Analytics Dashboard", "category": "Data", "stats": "9,026 Files", "description": "Full-stack data platform analyzing India's UPI payments across 788 districts with medallion architecture and 11-tab Streamlit dashboard.", "detailed_description": [ "Processed 9,026 JSON files from 3 sources across 788 districts/42 months using a 4-stage medallion pipeline into a DuckDB star schema", "Quantified market duopoly via HHI, measured geographic inequality using Gini coefficients, segmented adoption with K-Means", "Forecasted UPI volumes using Prophet/ARIMA; delivered 68 visualizations, validated by 83 tests via GitHub Actions" ], "tech": ["Python", "DuckDB", "Pandas", "Plotly", "Streamlit", "Power BI", "Prophet", "ARIMA", "Scikit-learn", "K-Means", "BeautifulSoup4", "PyArrow", "GitHub Actions"], "github": "https://github.com/prasad-gade05/UPI_DS_Project", "demo": "https://upi-analytics.streamlit.app/", "dataset": "https://huggingface.co/datasets/prasad-gade05/india-upi-ecosystem-2018-2025", "kaggle_dataset": "https://www.kaggle.com/datasets/prasadgade/india-upi-ecosystem-2018-2025" }, { "title": "Lex Simulacra", "subtitle": "AI Legal Courtroom Simulator", "category": "AI", "stats": "8 AI Agents", "description": "Multi-agent legal simulation with LangGraph for realistic trial proceedings.", "detailed_description": [ "Built an 8-agent legal simulation (defense, prosecutor, judge) achieving 99% citation accuracy via LangGraph workflow", "Reduced legal research time by 75% across 100K+ documents with RAG pipeline and BM25 re-ranking", "Achieved 3.2x context relevance gain and 60% engagement boost via context compression and hallucination detection" ], "tech": ["Python", "LangChain", "LangGraph", "FastAPI", "Streamlit", "ChromaDB", "Sentence-Transformers", "Ollama LLM", "Kanoon API", "BM25", "NLTK", "Scikit-learn"], "github": "https://github.com/prasad-gade05/Law_Courtroom_Simulator", "demo": null }, { "title": "Intrusion Detection", "subtitle": "TII-SSRC-23 Analysis", "category": "ML", "stats": "100% recall", "description": "Compared 6 ML/DL models on 8.6M samples. Tree models outperformed.", "detailed_description": [ "Achieved 99.99% recall and 100% precision benchmarking 6 models on the 8.6M-sample TII-SSRC-23 dataset (79 features)", "Tree models were 12.8x faster with 20x inference throughput and full SHAP interpretability", "Reduced classification errors by 30% via hybrid SMOTE-undersampling for 1:6653 class imbalance" ], "tech": ["Python", "Scikit-learn", "XGBoost", "PyTorch", "TensorFlow", "SHAP", "Pandas", "NumPy", "Imbalanced-learn"], "github": "https://github.com/prasad-gade05/IDS_on_TII-SSRC-23", "demo": null }, { "title": "KindHearts", "subtitle": "Donation Platform", "category": "Web", "badge": "1st Place", "stats": "80+ teams", "description": "Multi-role donation platform with crypto payments and real-time tracking.", "detailed_description": [ "Streamlined donation workflows across institutes, donors, and vendors with 60% reduction in request processing time", "Integrated real-time request tracking, crypto-based payments via MetaMask, and impact reporting (45% engagement boost)", "Implemented Kanban-style system for shopkeepers and comprehensive admin controls" ], "tech": ["React", "TypeScript", "MongoDB", "Vite", "Tailwind CSS", "React Router", "ESLint", "Context API"], "github": "https://github.com/prasad-gade05/KindHearts-Multi-Role-Donation-Management-Platform", "demo": null }, { "title": "Celestial Classifier", "subtitle": "SDSS Dataset Analysis", "category": "ML", "stats": "4 ML Models", "description": "Classification of celestial objects from Sloan Digital Sky Survey dataset.", "detailed_description": [ "Achieved 99.07% accuracy classifying 10,000 celestial objects into stars, galaxies, and quasars", "Improved Neural Network accuracy from 95.13% to 98.78% using Keras Tuner across 10 trials", "Benchmarked 4 classifiers with 5-fold CV and 960+ fits" ], "tech": ["Python", "Scikit-learn", "Keras", "Keras Tuner", "Pandas", "NumPy", "Matplotlib"], "github": "https://github.com/prasad-gade05/Celestial-Object-Classifier-using-Solana-Digital-Sky-Survey-Dataset", "demo": null }, { "title": "NutriSnap", "subtitle": "AI Nutrition Tracker", "category": "AI", "stats": "Image Analysis", "description": "Nutrition tracking using Gemini API for food image analysis.", "detailed_description": [ "Integrated Gemini API to analyze food images or manual entries with 95% accurate nutrient estimates", "Implemented daily calorie/macro goals and interactive dashboards", "Dual input modes (image/manual) with local SQLite storage" ], "tech": ["Flask", "Python", "HTML", "CSS", "JavaScript", "Gemini API", "SQLite", "Pandas", "OpenCV"], "github": "https://github.com/prasad-gade05/nutrition_tracker", "demo": "https://prasadsnutritiontracker.netlify.app/" }, { "title": "Audio Visualizer", "subtitle": "Real-Time Music Viz", "category": "Web", "stats": "3D Globe", "description": "Audio visualization with file upload, system capture, and 3D globe mode.", "detailed_description": [ "Real-time audio visualization with multiple dynamic modes (bars, radial waves, scatter plots)", "Fully client-side on GitHub Pages with file upload and microphone input", "1000+ unique visitors in the first month" ], "tech": ["TypeScript", "Web Audio API", "HTML5 Canvas", "shadcn/ui", "CSS3"], "github": "https://github.com/prasad-gade05/audio_visualizer_app", "demo": "https://prasadgade.dev/audio_visualizer_app/" }, { "title": "Attendance Tracker", "subtitle": "Smart Schedule & Attendance", "category": "Web", "stats": "Offline Ready", "description": "Attendance simulation to calculate required classes with goal tracking.", "detailed_description": [ "Subject-wise reports, student-wise analysis, editable records", "Offline-first using IndexedDB (Dexie.js) for persistent data", "Import/export records and data backup support" ], "tech": ["React", "TypeScript", "Vite", "Tailwind CSS", "Zustand", "Dexie.js", "Recharts"], "github": "https://github.com/prasad-gade05/attendance", "demo": "https://prasadgade.dev/attendance/" }, { "title": "Habit Tracker", "subtitle": "Privacy-First Tracking", "category": "Web", "stats": "Local Storage", "description": "Client-side habit tracking with GitHub-style contribution charts.", "detailed_description": [ "GitHub-style contribution charts, streaks, calendar view, and analytics", "IndexedDB via Dexie.js for persistent offline storage, no server dependency", "Modern UI with shadcn/ui, Tailwind, and Zustand state management" ], "tech": ["React", "TypeScript", "Vite", "Tailwind CSS", "shadcn/ui", "Zustand", "Dexie.js", "Recharts", "date-fns"], "github": "https://github.com/prasad-gade05/Habit-Tracker", "demo": "https://prasadgade.dev/Habit-Tracker/" }, { "title": "Portfolio", "subtitle": "Personal Developer Portfolio", "category": "Web", "stats": "4 Themes", "description": "Interactive portfolio with multi-theme support, 3D elements, and responsive design.", "detailed_description": [ "4 visual themes with auto-detect device preference", "3D Minecraft skin viewer (skinview3d + Three.js), paper physics playground", "Hidden easter eggs (Konami code, hesoyam), responsive for all screen sizes" ], "tech": ["React", "Three.js", "Framer Motion", "Vite"], "github": "https://github.com/prasad-gade05/prasad-gade05.github.io", "demo": "https://prasadgade.dev" } ] } }