Getting Started with Capella Datascience

Capella Datascience equips Eclipse with native capabilities for analyzing, visualizing, and manipulating complex multi-dimensional scientific data formats (NetCDF-3, NetCDF-4, HDF5, GRIB, CDF) alongside Model-Based Systems Engineering (MBSE) models in Eclipse Capella.

Key Capabilities at a Glance

• Unified Scientific Data Support: Open and inspect local (.nc, .ncml, .nc3, .nc4, .cdf, .hdf5) and remote datasets via OpenDAP (http://, https://) or AWS S3 (cdms3://, s3://).
• NcML Digital Twin Architecture: Automatically construct an EMF-based NcML 2.2 model reflecting dataset structures, enabling metadata manipulation, attribute overriding, and fast querying without re-writing binary data.
• Interactive Multi-Dimensional Visualization: Render high-performance 2D contour maps and heatmaps in NetCDFPlotView and NCEditor with customizable color palettes (Spectral, Jet, Plasma, Viridis, etc.), coordinate projections, and real-time slice navigation across extra dimensions (time, elevation, pressure levels, ensembles).
• Persistent Visualization Configurations (.ncconfig): Save, share, and reload complete visualization specifications (including colorbars, grid sections, projection CRS, labels, and slice indices) with automatic dataset proxy resolution.
• Geospatial & Apache SIS Integration: Reproject spatial data on-the-fly using Apache SIS across global and regional Coordinate Reference Systems (e.g. EPSG:4326, EPSG:3857).
• Python & Jupyter Workspace Integration: Execute data science routines in Jupyter notebooks and Python environments directly within the Eclipse workbench.

First Steps

1. Open the Data Exploration Perspective: Navigate to Window > Perspective > Open Perspective > Other... > Data Exploration.
2. Explore NetCDF Data: Follow the NetCDF Data Exploratory Intro to open and view your first NetCDF file.
3. Connect to THREDDS Catalogs: Follow the THREDDS Introduction to browse remote scientific servers.
4. Automate with Jupyter / Python: Follow the Jupyter Kernels Intro to interact with scientific data programmatically.