Hydroinformatics Lab, Tulane University

Visual Hydrology
Workflows, Powered by AI

Drag-and-drop hydrology analysis pipelines with native Python, R, and JavaScript execution. Connect to 50+ live data sources, run spatial analysis, and visualize results — all in one canvas. Includes an integrated AI assistant for code generation and analysis.

Run in Browser Desktop App
USGS Streamflow
Python LP3
IDF Curves
Results

Full-Stack Hydrology Platform

From raw data retrieval to spatiotemporal analysis, simulation, and AI-assisted coding.

Visual DAG Canvas

Build pipelines on an infinite canvas. Blocks execute in topological order. Click Run All or Run From Here. Per-block real-time status, stdout preview, and execution time are shown inline.

Native Python & R

The desktop app runs Python/R as native subprocesses — full access to pandas, numpy, scipy, rasterio, geopandas, xarray, and any conda/pip package. Browser falls back to Pyodide and WebR for zero-install use.

AI Code & Workflow Assistant

Describe your analysis in plain English — the AI writes the code block, picks data sources, and proposes visualization blocks. Supports OpenAI GPT-4o, local Ollama, LM Studio, and in-browser WebLLM. Context-aware: reads the live canvas structure.

50+ Live Data Sources

Each data block connects directly to a live API — no code needed. Sources include USGS NWIS, NOAA, NWM, CHIRPS, NASA POWER, NLDAS, MRMS, AORC, 3DEP, NHD+, WQP, HydroShare, CUAHSI, ECMWF ERA5, PRISM, HRRR, FEMA, Open-Meteo, ISRIC, World Bank, and 30+ more.

Geospatial Analysis

Leaflet map blocks render GeoJSON, shapefiles, and GeoTIFF rasters inline. Geoprocessing blocks: bbox clip, watershed delineation (NLDI), spatial join, zonal statistics, raster reprojection. Outputs export as GeoTIFF or Shapefile ZIP.

In-Place Visualization

Charts (line, bar, scatter, heatmap, box), maps (choropleth, raster overlay, marker cluster), and interactive tables render inside each result block. The Data Explorer tree lets you bind any JSON field as a chart axis with one click.

Parallel Map & Control Flow

Fan-out data across N values (years, stations, return periods) with Parallel Map. Fan-in with Aggregator blocks using concat, array-wrap, or object-merge. Conditional routing and convergence-loop blocks support calibration workflows.

Persistent Data Store

Every result saves automatically — IndexedDB in browser, SQLite on desktop. Browse, preview, and re-inject any saved dataset as a static canvas node. Binary outputs (GeoTIFF, NetCDF, HDF5) detected by magic bytes and saved correctly. Bulk ZIP export.

HydroShare Integration

Browse public resources by keyword or bounding box. Authenticate to access private resources. Upload result datasets as new HydroShare resources or append to existing ones — directly from the canvas block.

Supported Data Sources

Pull live data directly into your workflow from over 50 services — no API key required for most.

USGS NWIS
NOAA
NWM / NextGen
NASA POWER
CHIRPS
NLDAS
MRMS Radar
3DEP Elevation
NHD+ / NLDI
WQP (Water Quality)
HydroShare
CUAHSI WaterOneFlow
Open-Meteo
ECMWF ERA5
PRISM Climate
HRRR Forecast
FEMA Flood Data
AORC Forcing
ISRIC Soil
World Bank

How It Works

From raw data to publication-ready results in minutes.

1

Add a Data Source Block

Drop a data block and select your source — USGS, NOAA, HydroShare, a CSV upload, or any of 50+ APIs. Data flows downstream automatically.

2

Build Your Analysis Pipeline

Chain transform, statistics, code, and geoprocessing blocks. Each block's output is the next block's input. Use AI to generate code blocks from a description.

3

Run the Workflow

Click Run. Blocks execute in topological order (respecting dependencies). Python and R run natively on desktop for full library access.

4

Explore Results & Export

Switch to Results mode. Charts, maps, tables, and rasters render inline. Download any dataset as JSON, GeoTIFF, NetCDF, or ZIP. Save the workflow as a .blox file to share or reuse.

Documentation

Comprehensive guides, data source references, code block APIs, and more.

Open Documentation

Get HydroBlox

Desktop App

Web App

No install. Runs in any modern browser. Python via Pyodide (numpy, pandas, scipy). Best for sharing and quick exploration.

Open Web App

Windows

Portable .exe — no install needed. Double-click and run directly. Uses your system Python and R for full native library access.

Download .exe

macOS

Universal DMG for Intel and Apple Silicon (macOS 12+). Connect your Homebrew or Conda environment.

Coming Soon

Linux

AppImage — no root required. Works on Ubuntu, Fedora, and most distributions. chmod +x and run.

Coming Soon