How to set an interactive Matplotlib backend

Every figure created by Matplotlib passes through a backend that combines a renderer with either a GUI event loop or a file-only canvas. A desktop script that must open a pan-and-zoom window needs a GUI backend supported by the active Python environment.

The TkAgg backend draws through Tkinter and works well for a small standalone verification script. Select it with matplotlib.use(“TkAgg”) before importing matplotlib.pyplot so the first figure is created with the intended canvas and event loop.

This method targets a standalone Python process in a local desktop session. Notebook kernels manage GUI integration differently, while SSH sessions, containers, and servers without a forwarded display need a file-only or notebook-aware backend instead.

1. Run the Tkinter self-test from the Python environment that runs Matplotlib.

   $ python -m tkinter

   A small window should show the installed Tcl/Tk version.

2. Close the Tkinter test window after its version appears.
3. Create interactive_backend.py with the TkAgg selection before the pyplot import.

   interactive_backend.py

   import matplotlib
    
   matplotlib.use("TkAgg")
    
   import matplotlib.pyplot as plt

4. Append the figure construction section to interactive_backend.py.

   fig, ax = plt.subplots(layout="constrained")
   ax.plot([1, 2, 3, 4], [2, 5, 3, 6], marker="o")
   ax.set_title("Interactive backend check")
   ax.set_xlabel("Run")
   ax.set_ylabel("Value")

5. Append the backend report and blocking GUI event loop to interactive_backend.py.

   print(f"backend: {matplotlib.get_backend()}", flush=True)
   print(f"canvas: {type(fig.canvas).__name__}", flush=True)
    
   plt.show()

   A blocking plt.show() keeps the GUI event loop responsive until the figure window closes.

6. Run the completed script from a local desktop session.

   $ python interactive_backend.py
   backend: TkAgg
   canvas: FigureCanvasTkAgg

   A process without access to a desktop display cannot open a TkAgg window. A local session or configured display forwarding is required; otherwise use the related headless or notebook backend.

7. Confirm the plot window displays the line and navigation toolbar.
8. Close the first plot window to end its blocking GUI event loop.
9. Rerun the saved interactive_backend.py script to confirm TkAgg opens a new interactive canvas.

   $ python interactive_backend.py
   backend: TkAgg
   canvas: FigureCanvasTkAgg