Notebooks are cells of logic. You could conceivably change the idea of notebook cells to be an instance of a function that points to raw data and returns raw data.
Perhaps this just Alteryx though
I'm picturing the ability to write a Python function with the parameters being just like the parameters in an Excel function. You can drag the cell and have it duplicated throughout a row, updating the parameters to correspond to the rows next to it.
It would exponentially expand the power of excel. I wouldn't be limited to horribly unmaintainable little Excel functions.
VBA can't be used to do that, can it? As far as I understand (and I haven't investigated VBA too much) VBA works on entire spreadsheets.
Essentially, replace the excel formula `=B3-B4` with a Python function `subtract(b3, b4)` where Subtract is defined somewhere more conveniently (in a worksheet wide function definition list?).
The ubiquity of Excel is both a blessing and a curse in that everyone has it, so everyone uses it, regardless of whether or not it is the best tool for the job.
as of now jupyter/ipython would not recompute `subtract(b3, b4)` if you change b3 or b4, this has positive and negative (reliance on hidden state and order of execution) effects.
I too would really like something like this, but I think it is pretty far away from where jupiter is now.
> Traitlets is a framework that lets Python classes have attributes with type checking, dynamically calculated default values, and ‘on change’ callbacks. https://traitlets.readthedocs.io/en/stable/
> Traitlet events. Widget properties are IPython traitlets and traitlets are eventful. To handle changes, the observe method of the widget can be used to register a callback https://ipywidgets.readthedocs.io/en/stable/examples/Widget%...
You can definitely build interactive notebooks with Jupyter Notebook and JupyterLab (and ipywidgets or Altair or HoloViews and Bokeh or Plotly for interactive data visualization).
> Qgrid is a Jupyter notebook widget which uses SlickGrid to render pandas DataFrames within a Jupyter notebook. This allows you to explore your DataFrames with intuitive scrolling, sorting, and filtering controls, as well as edit your DataFrames by double clicking cells. https://github.com/quantopian/qgrid
Qgrid's API includes event handler registration: https://qgrid.readthedocs.io/en/latest/
> neuron is a robust application that seamlessly combines the power of Visual Studio Code with the interactivity of Jupyter Notebook. https://marketplace.visualstudio.com/items?itemName=neuron.n...
"Excel team considering Python as scripting language: asking for feedback" (2017) https://news.ycombinator.com/item?id=15927132
OpenOffice Calc ships with Python 2.7 support: https://wiki.openoffice.org/wiki/Python
Procedural scripts written in a general purpose language with named variables (with no UI input except for chart design and persisted parameter changes) are reproducible.
What's a good way to review all of the formulas and VBA and/or Python and data ETL in a spreadsheet?
Is there a way to record a reproducible data transformation script from a sequence of GUI interactions in e.g. OpenRefine or similar?
"Within the Python context, a Python OpenRefine client allows a user to script interactions within a Jupyter notebook against an OpenRefine application instance, essentially as a headless service (although workflows are possible where both notebook-scripted and live interactions take place. https://github.com/OpenRefine/OpenRefine/wiki/Jupyter
Are there data wrangling workflows that are supported by OpenRefine but not Pandas, Dask, or Vaex?
Not sure how they can converge.
This kind of thing exists at a larger scale for pipeline visualization. I could see it working for notebooks.