Add basic plot of length vs. frequency for line-based sensors

[bwalraven]

These plots would be useful to view the distribution of frequency and length of an entire network. It relies on xarray.DataArray in the OPENSENSE format.

One (general) question is how much aesthetic details/choices to add to the function. Now only axis labels are provided by default, and the possibility to add a grid (boolean) is added as an option.

[bwalraven]

• [ ] Add margin plots that show distribution as bar graph Do we add this as option in basic scatter plot? The function would return 3 axes objects (1 for each graph), rather than 1 object in the basic plot. Also do we use seaborn (easier with jointplots but extra dependency) or create extra axes in matplotlib?
• [ ] Add scatter density plot. Probably this is best done as a separate function. Though the code would be mostly redundant, this returns a PolyCollection rather than a PathCollection and requires additional arguments like colormap and min_count.
• [ ] Add distribution of frequency and length, both in percentages, and as total numbers (in which case length would be plotted unnecessarily double for all the sub-links, but that way it matches the numbers in frequency distrubtion plots)
• [ ] Add distribution of polarization: HH, VV, or HV for each CML
• [ ] Add number of link paths and sublinks during the time period we have data.
• [ ] Add average availability throughout the day.
• [ ] Optionally add orientation.
• [x] Add examples of all plots to notebook