A new tack I’m taking this semester is to remove the use of Pandas from my Physical Chemistry Lab course. Pandas is an excellent tool with many evangelists. After some careful examination (and questions from a colleague), I realized that my reasons for using it the course weren’t justified. For the most part, I was using Pandas because it made pretty outputs in a Jupyter notebook… hardly a good reason for introducing a new tool to students already overwhelmed with new tools. In an effort to teach the basics (in this case, Python with NumPy, SciPy, and Matplotlib), I decided to replace my Pandas calls with their more fundamental equivalents. None of this is to day Pandas isn’t an excellent tool, only that I didn’t want to use very limited course time to teach an additional tool (and introduce more sources of confusion).
One of my frequent calls to Pandas is in importing a series of spectral data files (UV/Vis absorption or fluorescence). These are tab- or comma-separated files with some sort of header. The header is of predictable length depending on the technique, so by skipping a few rows, the imported data are essentially columns of wavelength and intensity values. In Pandas, the read_csv() function is used to read in data, even though it can read in data that isn’t comma-separated (pet peeve number one with Pandas). Columnar data can be appended to a Pandas array very easily using the same syntax used to make a new dict item (dict[new_key] = new_value). This ability to easily create new columns is handy, as data transformations are often necessary. This can, of course, be done using simple NumPy arrays, but I haven’t figured out how to add a column to a structured array.
I’ll update this post once I have more to say.