Include the scripts to generate the figures

mechmotum / fall-probability-paper

Paper on the effect of the balance assist system on fall probability.

https://mechmotum.github.io/research/balance-assist-bicycle.html

1 stars 0 forks source link

Include the scripts to generate the figures #1

Closed moorepants closed 2 months ago

moorepants commented 7 months ago

It would be nice if we include the scripts that generate the stats results tables and the figures here. I typically make a src/ directory and put all those scripts there which generates figures and tables to a figures/ and tables/ directory. You can see an example here: https://github.com/csu-hmc/perturbed-data-paper. It is best if we DO NOT commit binary files (PDF, png, etc) to the repo (at least at first). For figures we hand draw, we have to commit them, but if it happens to be svg, it is preferred to commit those.

For the data, we could upload the processed csv files here in a data/ folder. These would be the Python or R Data Frames that have all the data used to produce the stats tables and figures. These are small text files so fine to just commit them and they can be generated in Marten's other repos (the ones that take all the raw data and process). I think that is simplest and least work for this paper repo.

moorepants commented 7 months ago

I created a src directory with figure scripts and a conda environment file, which now all builds on the github action.

mthaitjema commented 6 months ago

To do:

[x] Create plots that show desired/actual force and torque on the handlebars with example data from experiments
[x] Time on x-axis relative to start of perturbation? Now it's time since start of data collection, but that's not really relevant for the plots.
[x] Create plots that show roll and steer angle after a perturbation
[ ] Auto run script from makefile
[ ] Auto generate tables
[x] Create probability plots
[ ] CSV instead of parquet?

moorepants commented 6 months ago

I see you are checking things off but no updates in the repository code. Please add these things as you go along so we can each work from the same version.

mthaitjema commented 6 months ago

I made a branch for this issue that I’m working in, it is linked to this issue

moorepants commented 6 months ago

With just us two, branches probably aren't needed, feel free to push to master. If you do make branches then a draft PR signals that there is something I can look at.

As soon as you get a new figure in, I can work on text about it. If you can put the few non-code generated figures in now, I can write about them. Otherwise I'm sorta waiting on things.

mthaitjema commented 6 months ago

Makes sense! I will merge the branch and work on the master from now on. I will push the non-generated images when I get home this afternoon 👍

moorepants commented 6 months ago

Great. Make sure to pull often, as it reduces chances of merge conflicts if I touch any files you are working on.

mthaitjema commented 6 months ago

If you want to generate the figures of the forces and of the roll and steer angle you can run the generate_time_series_imgs.py file. It should create a imgs of folder and generates the plots for the data of one participant

moorepants commented 6 months ago

I tried it already, works perfectly! Note that I moved images to figures/ (directory name I had already chosen).

moorepants commented 2 months ago

We have the main scripts in place.