Layout for playground concept graph

[mmaelicke]

@cojacoo,

I quckly drafted the concept graph figure for learning about Knightian uncertainty. I basically rebuilts what I have seen on your presentation. It is meant to be the starting point into the actual application to explain the concept.

You can run that app by checking out the branch and locally install the ruins package

git checkout uncertaintry_draft
pip install -e .

And then run within the repo

streamlit run ruins/apps/uncertainty.py

Is the general layout and flow through this example what you had in mind? Obviously, the labels and text need some replacements and a little bit of work is needed for the figure, but this draft should point into the right direction.

[codecov[bot]]

Codecov Report

Merging #67 (c543755) into main (06ad9d7) will decrease coverage by 1.00%. The diff coverage is 0.00%.

@@            Coverage Diff             @@
##             main      #67      +/-   ##
==========================================
- Coverage   20.09%   19.09%   -1.01%     
==========================================
  Files          19       19              
  Lines        1612     1697      +85     
==========================================
  Hits          324      324              
- Misses       1288     1373      +85     

Impacted Files	Coverage Δ
ruins/apps/uncertainty.py	0.00% <0.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 06ad9d7...c543755. Read the comment docs.

[cojacoo]

Very nice. I can easily hook into this, because the fun starts when the distributions are skewed and when there are some threshold processes changing your expected outcome. I will (so time permits) use your streamlit app to develop it further. However: I have an issue with git due to the LFS quota. If I clone the repo, it does not allow me to get all the branches... I would be happy to avoid any conflicting versions...

[mmaelicke]

Skewed distributions are interesting. Then, I will make the used PDF function a changeable parameter for future developments. As of now, it is hard-coded to scipy.stats.norm.pdf. In the future, we can dynamically change it to any PDF supported by scipy. There are dozens of continuous distributions supporting the .pdf method: https://docs.scipy.org/doc/scipy/reference/stats.html#continuous-distributions

We can basically use any of them. We just need to keep in mind that they don't come with the same function signature, but I think we can limit ourselves to maybe 3 or 4 PDFs.