A minimal GUI for a quick view of netCDF files. Aiming to be a drop-in replacement for ncview and panoply.

ncvue is a minimal GUI for a quick view of netCDF files. It is aiming to be a drop-in replacement for ncview and panoply, being slightly more general than ncview targeting maps but providing animations, zooming and panning capabilities unlike panoply. If ncvue is used with maps, it supports mostly structured grids, more precisely the grids supported by cartopy.

ncvue is a Python script that can be called from within Python or as a command line tool. It is not supposed to produce publication-ready plots but rather provide a quick overview of the netCDF file(s).

The complete documentation for ncvue is available from:

https://mcuntz.github.io/ncvue/

ncvue can be run from the command line:

ncvue netcdf_file.nc
ncvue netcdf_file1.nc netcdf_file2.nc

or from within Python:

from ncvue import ncvue
ncvue(['netcdf_file.nc'])

where the netCDF file is optional. The latter can also be left out and a netCDF file can be opened with the "Open File" button from within ncvue. The netCDF has to be given in a list because several netcdf files can be given as in the second example from the command line.

Note, ncvue uses the TkAgg backend of matplotlib. It must be called before any other call to matplotlib. This also means that you cannot launch it from within iPython if it was launched with --pylab. It can be called from within a standard iPython, though, or using ipython --gui tk.

We also provide standalone (Apple notarized) macOS applications that come with everything needed to run ncvue including Python:

• ncvue 4.0 for macOS 10.x with Intel processor
• ncvue 5.1 (Aqua, Intel) for macOS 14 [Sonoma] in Aqua look on Intel processors
• ncvue 5.1 (Aqua, ARM) for macOS 14 [Sonoma] in Aqua look on ARM processors
• ncvue 5.1 (CustomTkinter, Intel) for macOS 14 [Sonoma] in CustomTkinter look on Intel processors
• ncvue 5.1 (CustomTkinter, ARM) for macOS 14 [Sonoma] in CustomTkinter look on ARM processors

ncvue 4.0 should work from macOS 10.13 (High Sierra) onward on Intel processors. ncvue > 5.0 is either for Intel processors or for Apple Silicon (ARM) chips. It comes in the standard Aqua look or uses the CustomTkinter UI-library.

On opening, ncvue presents three panels for different plotting types: Scatter or Line plots, Contour plots, and Maps. This is the look in macOS light mode (higher resolution images can be found in the documentation):

All three panes are organised in this fashion: the plotting canvas, the Matplotlib navigation toolbar and the pane, where one can choose the plotting variables and dimensions, as well as plotting options. You can always choose another panel on top, and open another, identical window for the same netCDF file with the button "New Window" on the top right.

If CustomTkinter is installed (python -m pip install customtkinter), then the general layout looks like this:

CustomTkinter does not work well with Conda.

If ncvue detects latitude and longitude variables with a size greater than 1, it opens the Map panel by default. This is the Map panel in macOS dark mode, describing all buttons, sliders, entry boxes, spinboxes, and menus:

If it happens that the detection of latitudes and longitudes did not work automatically, you can choose the correct variables manually. Or you might use the empty entries on top of the dropdown menus of the latitudes and longitudes, which uses the index and one can hence display the matrix within the netCDF file. You might want to switch of the coastlines in this case.

You might want to switch off the automatically detected "global" option sometimes if your data is on a rotated grid or excludes some regions such as below minus -60 °S.

All dimensions can be set from 0 to the size of the dimension-1, to "all", or to any of the arithmetic operators "mean", "std" (standard deviation), "min", "max", "ptp" (point-to-point amplitude, i.e. max-min), "sum", "median", "var" (variance).

Be aware that the underlying cartopy/matplotlib may (or may not) need a long time to plot the data (with the pseudocolor 'mesh' option) if you change the central longitude of the projection from the central longitude of your data, which is automatically detected if "central lon" is set to None. Setting "central lon" to the central longitude of the input data normally eliminates the problem.

The mapping window looks like this with CustomTkinter:

If ncvue does not detect latitude and longitude variables with a size greater than 1, it opens the Scatter/Line panel by default. This is the Scatter/Line panel in macOS dark mode (Aqua look), describing all buttons, sliders, entry boxes, spinboxes, and menus:

The default plot is a line plot with solid lines (line style 'ls' is '-'). One can set line style 'ls' to None and set a marker symbol, e.g. 'o' for circles, to get a scatter plot. A large variety of line styles, marker symbols and color notations are supported.

ncvue builds automatically a datetime variable from the time axis. This is correctly interpreted by the underlying Matplotlib also when zooming into or panning the axes. But it is also much slower than using the index. Selecting the empty entry on top of the dropdown menu for x uses the index for the x-axis and is very fast. Plotting a line plot with 52608 time points takes about 2.2 s on my Macbook Pro using the datetime variable and about 0.3 s using the index (i.e. empty x-variable). This is especially true if one plots multiple lines with 'all' entries from a specific dimension. Plotting all 10 depths of soil water content for the 52608 time points, as in the example below, takes also about 0.3 s if using the index as x-variable but more than 11.1 s when using the datetime variable.

The last panel provide by ncvue draws contour plots. This is the Contour panel in macOS dark mode, describing all buttons, sliders, entry boxes, spinboxes, and menus:

This produces also either pseudocolor plots ('mesh' ticked) or filled contour plots ('mesh' unticked) just as the Map panel but without any map projection.

ncvue is an application written in Python. If you have Python installed, then the best is to install ncvue within the Python universe. The easiest way to install ncvue is thence via pip if you have cartopy installed already:

pip install ncvue

Cartopy can, however, be more elaborate to install. The easiest way to install Cartopy or directly ncvue is by using Conda. After installing, for example, Miniconda:

conda install -c conda-forge ncvue

See the installation instructions in the documentation for more information on installing Cartopy and ncvue with pip.

ncvue is distributed under the MIT License. See the LICENSE file for details.

Copyright (c) 2020-2024 Matthias Cuntz

ncvue uses CustomTkinter if installed. Otherwise it uses the Azure 2.0 theme by rdbende on Linux and Windows.

The project structure of ncvue was very originally based on a template provided by Sebastian Müller but has evolved considerably since.

Different netCDF test files were provided by Juliane Mai.