Recently I installed Python 3.8 and subsequently had some difficulty installing the scientific packages I needed (SciPy, Pandas, Matplotlib etc.). Poked around and found a really excellent source of up-to-date Windows binary files of many open-source scientific packages. Here's the nitty-gritty info.

As part of the work I do for the OU as an AL and also as bait of a consultant on developing Python content (sounds grander than it is), I wanted to get into the intricacies of using Matplotlib to produce acceptable sub-plots for scientific reports. Here's my attempt which covers using common x axis labels, overall x and y axis label text and 2 or more column plots. It deals with plot numbers that don't 'fill up' a nice rectangle.

Here's an interesting little function I cooked up. With many experimental spectra, there is an underlying 'background' signal that distorts the actual useful signal. Sometimes you can remove this by taking background spectra from somewhere where the actual signal is absent and subtract this (usually an average of several of them) from the useful spectrum.

In some cases, this isn't available, but the background is easy to simulate as, for example, it may be a straight line 'slope'. hence a simple function can be 'fitted' and subtracted.

However, there are clearly some cases where neither of these methodologies are possible. This post gives a simple function that I thought up and have used successfully.

Here is the Python I've used to generate the plots used on the COVID-19 pages (well it will be in time)

As part of a series, I thought it would be good to do some slightly advanced or more obscure tips and tricks for Python. Here I'm looking at Pandas.