Working with Matplotlib

Matplotlib is a comprehensive library for creating static, animated, and interactive visualizations in Python.

See in Neptune  Code examples  

Before you start

• Set up Neptune. Instructions:
  • Installing Neptune
  • Setting your API token
  • Setting the project name
  • Creating a Neptune project

• Install Matplotlib:

  pipconda

  pip install -U matplotlib
  

  conda install -c conda-forge matplotlib
  

(Optional) To log static Matplotlib figures as interactive Plotly charts, install the plotly library:

pipconda

pip install -U plotly

conda install -c conda-forge plotly

Plotly incompatibility

Plotly is not compatible with the latest Matplotlib (3.5.0+) due to the fact that mpl_to_plotly() uses deprecated Matplotlib functionalities.

To use Plotly together with Matplotlib, you may need to downgrade your Matplotlib version to <3.5. For more information, see Plotly issue 1568 and issue 3624 on GitHub.

Matplotlib logging example

1. Import Neptune and start a run:

   import neptune
   
   run = neptune.init_run()  # (1)!
   

   1. If you haven't set up your credentials, you can log anonymously: neptune.init_run(api_token=neptune.ANONYMOUS_API_TOKEN, project="common/matplotlib-support")

2. Create a sample figure:

   import matplotlib.pyplot as plt
   import numpy as np
   
   np.random. Seed(42)
   data = np.random.randn(2, 100)
   
   figure, ax = plt.subplots(2, 2, figsize=(5, 5))
   ax[0, 0].hist(data[0])
   ax[1, 0].scatter(data[0], data[1])
   ax[0, 1].plot(data[0], data[1])
   

3. Log the figure to Neptune

   1. As a static image:

      run["static-img"].upload(figure)
      

   2. (With Plotly installed) As an interactive Plotly chart:

      from neptune.types import File
      
      run["interactive-img"].upload(File.as_html(figure))
      

      Info

      Not all Matplotlib charts can be converted to interactive Plotly charts. If the conversion is not possible, Neptune falls back to logging the chart as an image.

4. To stop the connection to Neptune and sync all data, call the stop() method:

   run.stop()
   

   Using stop() is especially important in Jupyter Notebook or other interactive sessions, as the connection otherwise remains open until the session ends completely.

5. To open the run, click the Neptune link in the console output.

   Example link: https://app.neptune.ai/common/matplotlib-support/e/MAT-1

6. Explore the figures in the All metadata section: