Joanna Diong

Note to self: [dowhy](https://github.com/microsoft/dowhy), another causal package, now supports [DAGitty](http://www.dagitty.net/dags.html) syntax to generate causal graphs (see [docs](https://microsoft.github.io/dowhy/example_notebooks/dowhy_simple_example.html)). I haven't looked under the hood yet, but their graphs are implemented a...

Cool find!: I just stumbled across the program `mermaid` that is a Java-based program that renders Markdown code into flowcharts and other diagrams. Looking at the flowchart features of nodes...

From the code above: > ```python > import networkx as nx > > dag > > fig = plt.figure(figsize=(6, 5)) > ax = plt.subplot(1, 1, 1) > positions = nx.spectral_layout(dag)...

Interesting, thanks! Did as you advised, and I managed to save the dag. This is a dag generated by DAGitty, and plotted using the output positions. But interestingly the orientation...

Thanks! Yes, I had supplied `dag` the positions from DAGitty (as dict of lists) in `dag.draw_dag(positions=pos)`. But it turns out that `nx.spectral_layout(dag.dag)` changes these positions: ```text original positions from DAGitty:...

From a previous comment on DAGitty positions, > A -> Z [pos="-0.791,-1.045"] > > Like what does the positivity after A -> Z do differently than the declared positions for...

Question. I've been building the plotting features in an external private repo, and wondering how to proceed with merging the functionality into `zEpid`. I felt it belongs better there, but...

Hi @MarcoGorelli, Many thanks for the suggestions. There are 2 traces in the chapter that call `trace_to_dataframe()`. **trace_4_1** I received an error for the first approach: ```python # Code 4.32...

Thanks @MarcoGorelli! **Code 4.32** ``` # Code 4.32 trace_df = pm_data.posterior.to_dataframe() trace_df.cov() ``` I used `pm` instead of `pm_data` because I couldn't find where `pm_data` was defined earlier? **Error at...

Got it, thanks! Fixed the structure of `trace_4_1` and `trace_4_3` to get dataframes this way: ```python trace_df = pm_data.posterior.to_dataframe() trace_df.cov() ``` It dealt with the long error. The notebook now...