Jupyter Progress Bars

Posted by James Gardiner on May 3, 2016

Jupyter notebooks have a number of built in widgets that can be used to enhance a notebooks functionality and appearance. One of these is the FloatProgress widget, which creates an updateable visual progress bar.


I was recently batch geocoding a large number of address strings inside a Jupyter notebook, and was looking for an easy visual way to track the progress of this operation, which is run as a pandas.Series.map() method.

My first thought was tqdm, “A Fast, Extensible Progress Meter” for Python, though from what I can see it doesn’t offer the ability to work with Panda’s Series objects (but it can be used with pandas.DataFrame.apply() methods) (edit - This has now been implemented. See https://github.com/tqdm/tqdm/issues/214 for details).

Not wanting to print and update a string, a bit of searching threw up this thread, the final comment of which shows how to implement a progress bar widget in Jupyter. The syntax in this comment is now deprecated but essentially all that is needed is to import the FloatProgress class from ipywidgets and the display method from the IPython.display module:

from ipywidgets import FloatProgress
from IPython.display import display

Then create a FloatProgress object with the relevant min and max values and display it below the selected cell:

f = FloatProgress(min=0, max=100)
display(f)

The bar can be updated by setting the f.value variable:

f.value += 1

Other keyword arguments for FloatProgress object include step which controls the step (obviously!), description which sets the bar’s name and bar_style which sets the color of the bar. More info can be found at the ipywidgets GitHub page.

This is a simple, easy and very quick way to get a progress bar up and running in Jupyter.