PhenoFit Pro is a modern, browser-based web application designed for advanced phenological data analysis. Developed to eliminate the complexity of R/Python scripts, this tool empowers researchers, scientists, and students to interactively explore, model, and extract phenological events from time-series vegetation index data.
🌐 Launch Now: https://itsakashpandey.github.io/PhenoFit-Pro/
PhenoFit Pro is a purpose-built web application, tailored to perform complex phenology tasks in an intuitive and interactive way. It’s built for those who analyze vegetation indices like NDVI, GCC, EVI, ExG, etc., and need precise control over curve fitting and visualization, all from within a browser.
- Loads data from CSV and Excel files (
.csv,.xlsx,.xls) - Intelligent column detection: auto-suggests X and Y based on common names like
date,doy,gcc,ndvi
- Supports Double Logistic, Single Logistic, and smoothing filters like LOESS, Moving Average, Savitzky-Golay
- Interactive slider-based tuning for parameters (a1 to a6)
- Lock/unlock parameters during optimization
- One-click numerical optimization for best fit
- Detects outliers using:
- Standard Deviation
- Interquartile Range (IQR)
- Moving Window SD
- Two-phase process: pending preview → confirmed removal
- Load external "grouping" files to overlay time windows (e.g. rainfall, drought) as:
- Vertical lines for events
- Shaded bands for periods
- Click-to-edit any element: point, curve, axis, background, labels, etc.
- Reposition draggable elements like legends, tooltips, and annotation boxes
- Intelligent Chat Interface: Interactive AI assistant with multiple personalities (Sane, Moody, Weirdo)
- Natural Language Commands: Control chart styling using plain English commands
- Context-Aware Responses: Maintains conversation context for follow-up commands
- Smart Entity Recognition: Understands chart elements, colors, shapes, and styling properties
- Multi-Model Support: Powered by advanced LLMs including DeepSeek R1, Qwen 2.5, and Gemini 2.0 Flash
- Request Cancellation: Stop button for LLM requests with full abort support
- Download high-res PNG chart images
- Export full results as Excel (.xlsx) with:
- Raw & fitted values
- Removed outliers
- Optimized parameters
- Fit statistics (R², RMSE)
- SOS, POS, EOS dates
| Layer | Technology |
|---|---|
| 🧱 UI Framework | React (with Hooks) |
| 🎨 Styling | Tailwind CSS (via CDN) |
| 📊 Charting | Recharts |
| 📄 File Parsing | Papaparse (CSV), SheetJS (Excel) |
| 📦 Modules | ES Modules + ImportMap (build-less) |
| 🖼 PNG Export | html-to-image |
| 🤖 AI Integration | OpenRouter API with multiple LLM models |
| 🧠 NLU Engine | Custom Natural Language Understanding |
| 💬 Context Management | Advanced dialogue state management |
This stack enables zero-install, zero-server usage—everything happens in-browser.
Akash Kumar
PhD Scholar, Geomatics Engineering, IIT Roorkee
📫 akash_k@ce.iitr.ac.in
🌐 itsakashpandey.github.io/PhenoFit-Pro
🔗 GitHub | LinkedIn
Released under CC BY-NC-ND 4.0. Attribution required. No commercial use or modification without permission.