PyQtGraph VS matplotlib

Image, 3D, or data visualization applications using OpenCV and the SciPy ecosystem. The Graphics View Framework can display an image and let the user interact with it, and the Python ecosystem is very rich for image processing, data analysis, and visualization. For example, LabelMe for image labeling, PyQtGraph for scientific graphics, or custom QWidget integration in Maya.

I did something like this recently but I used pyqtgraph: https://www.pyqtgraph.org/

Most of the issues are with GPU support. Some libraries still struggle with it AFAIU but I don't use GPU much. M1 native is mostly good everywhere, except Pyqt + Qt 5 (Qt6 is native, but some libraries don't support it yet), But for those you can just start a bash shell under intel emulation and all is fine. Performances do take a hit take hit, but are still better in M1+emulation intel than native on my old mac.

And for those who would need Python and Qt "complexified" :

https://www.pyqtgraph.org/

> PyQtGraph is a pure-python graphics and GUI library built on PyQt / PySide and numpy. It is intended for use in mathematics / scientific / engineering applications. Despite being written entirely in python, the library is very fast due to its heavy leverage of NumPy for number crunching and Qt's GraphicsView framework for fast display. PyQtGraph is distributed under the MIT open-source license.

[...]

> Packaging for Distribution

Very impressive! Have you looked into pyqtgraph? It's based on the widely used PyQt/PySide packages. Using the included video speed test, I am able to process a 512 x 512 picture at nearly 1000 fps (though my monitor is capped at 144 Hz) with an RTX 2080 Ti. I encourage you to open up the examples and see what capabilities it has for your system.

If you decide you like widgets, check out https://www.pyqtgraph.org/

pyqtgraph.Way faster than matplotlib.

Matplotlib: Matplotlib is a 2D plotting library that allows you to create visualizations of your data. It's a powerful tool for data analysis, but the syntax can be complex and the customization options can be overwhelming. GitHub - https://github.com/matplotlib/matplotlib

I will think more about what I want to say next, but for now, I would like to say that I need the super-particles and PIC methods as I think that is the way forward for me. Are there ways to implement these methods in matplotlib, Visit or Paraview? Do I take existing code and import it into those programs to visualize it? Or can I directly program/simulate something in those visualizion tools without needing to import any code?

Your choices are an n-body simulation (e.g., LAMMPS) with Coulomb interactions or, if your electrons are sufficiently sparse, a particle-in-cell (e.g., Starfish). Your best bets for visualization are going to be matplotlib or something more user-friendly like Visit or Paraview. Without a neutralizing background, however, your electrons are just going to repel each other, hit the walls, and disappear - there's not going to be much interesting to visualize. What are you actually trying to simulate? With more information maybe you could receive some more targeted advice.

There are plenty of data visualization tools in python, but probably the easiest to get started with is Matplotlib

An example of how I would do this is to just plot your data on a line graph (https://matplotlib.org/) . Are there any repeating trends? Next try splitting your data into day of the week, day of the month, months, etc. Look for any kind of seasonality (we're trying to use the past to predict the future, so if the future is not like the past our models will fail).

Install libraries and tell matplotlib to plot inline (inside notebook) with the help of % magic functions which sets the backend of matplotlib to the inline backend:

The thing is, you will also need to go get the wheels (or *.tar.gz sources) for all of the dependencies of your packages as well! Over in matplotlib's setup.py you can find the following:

GitHub history for rcsetup.py

This commit. You can get that by clicking "View Git Blame" on line 307.