Applied math and a spreadsheet-like environment¶
This is an example of a fairly complex computation that benefits from a spreadsheet-like capability.
We want to change the measurements, and recompute a very complex estimation.
This use case of “enter data – see resulting computation” works nicely for spreadsheets where the computations aren’t too tricky. We can examine them for correctness, and trust they are working.
When the computations become non-trivial, a spreadsheet can be challenging to validate and debug. We may not be able to simply examine a computation cell to be sure it’s right.
This is where Jupyter Lab can be a dramatic improvement over a spreadsheet. In these examples we’ll do complex symbolic math (“algebra”) as well as ordinary applied math to compute useful values.
Components Required¶
You’ll need Python, of course. You may want to use Miniconda to help install Python.
https://docs.conda.io/en/latest/miniconda.html
With miniconda, you can build your Python environment.
conda create --name model python=3.9
conda activate model
Once you have Python, you can then install the suite of tools for working with more advanced symbolic math, Sympy https://docs.sympy.org/latest/index.html
python -m pip install sympy jupyter-lab
With the Jupyter lab you can launch the .ipynb notebooks to
make changes, and do computations. Like a spreadsheet. But
with real code and real math.
The final publication of a report or book can be done with two additional tools.
Jupyter {Book} https://jupyterbook.org/intro.html
python -m pip install jupyter-book
Requires MyST{NB} https://myst-nb.readthedocs.io/en/latest/
python -m pip install myst-nb
Document Production¶
Run the following:
jupyter-book build .
cp _build/html docs
This will create the HTML version.
To make a PDF, use the following.
jupyter-book build . --builder pdflatex
(This generally relies on having TeX installed.)
Fair warning: the structure used in this first release doesn’t look very good in a PDF. An outline that’s accetable in HTML can turn out to be weirdly structured when viewed as a PDF.