The focus of my summer internship was working on MetPy, both adding new
functionality and updating existing documentation to reflect recent updates to MetPy. In
particular, I focused on expanding current declarative syntax functionality to Matplotlib’s
pcolormesh via the RasterPlot class. Declarative syntax enables quick and easy plotting
of data, particularly for users new to Python or MetPy.
Currently, users can use declarative syntax to make plots that use the Matplotlib
contourf(). Though these functions are able to make a wide
range of plots, using
pcolormesh() declaratively was missing. Given that
pcolormesh() is commonly used to create raster plots, radar plots, etc., we found it
important to add declarative syntax support to MetPy.
With declarative syntax (shown at left), users do not have to call the
pcolormesh() and can simply pass through the data, field, and
level the user would like to plot. The Matplotlib calls would happen in the backend
and output the image to the right. To do this, we followed the software development
process of planning, designing, implementing, testing, and integrating. I found this
to be particularly exciting since this was my first time going through the
development process on a piece of code that was to be released and utilized by users.
Outside of adding declarative syntax for raster plots, I also worked on updating,
improving, and adding example documentation for radar. Some functions that were
tagged in past releases were not included in the radar documentation and I worked to
make sure functions like
azimuth_range_to_lat_lon that MetPy has were used in
tutorials for new or beginning users to MetPy.
I could not have done all of this work without the support of Drew Camron, Ryan May, and the entire Unidata staff. I’d also like to thank fellow Unidata interns Hassan and Rhoen, and the other NCAR/UCAR/UCP and SOARS interns for a great summer together in Boulder!