0

There is an easy way to generate word frequency plots with NLTK:

my_plot = nltk.FreqDist(some_dynamically_changing_list[0]).plot(20)

some_dynamically_changing_list contains a list of texts that I want to see the word frequency for. The resulting my_plot is an AxesSubplot object. I want to be able to take this object and directly insert it to a dynamically sized subplotted grid. The closest I've gotten to the solution after scouring SO and Google is this. However, I have an issue because I don't want to use plt's plot function. So for example I'd have a dynamic subplot grid:

import matplotlib.pyplot as plt

tot = len(some_dynamically_changing_list)
cols = 5
rows = tot // cols 

if tot % cols != 0:
    rows += 1
    
position = range(1, tot + 1)

fig = plt.figure()
for k in range(tot):
    ax = fig.add_subplot(rows, cols, position[k])

    my_plot = nltk.FreqDist(some_dynamically_changing_list[k]).plot(20)
    ax.plot(my_plot) #This is where I have my issue.

Of course this won't work, as ax.plot expects data, not a plot. But I want to plug my my_plot into that specific slot. How can I do that (if I can)?

lte__
  • 7,175
  • 25
  • 74
  • 131
  • You can't "take" a subplot and place it somewhere. When creating the plot, it already needs to be on the correct spot. Most functions that create plots allow an `ax=` parameter, so maybe in this case you could try `nltk.FreqDist(...).plot(..., ax=ax)`? Further, most such functions create the plot on the "current subplot", which in your case would be already set via `fig.add_subplot`. Are you sure you are using the latest github version of `nltk`? – JohanC Oct 05 '22 at 15:43
  • Looking into the [source code]((https://github.com/nltk/nltk/blob/f989fe65d421e7ea4d1037a00f07eaeee3ad6a29/nltk/probability.py#L247-L249), `nltk`'s `plot` doesn't support an `ax=` parameter, but does use `plt.gca()` to make sure it draws on the "current ax" (the currently active subplot). Anyway, you should remove `ax.plot(my_plot)` as such a thing doesn't make sense in matplotlib. – JohanC Oct 05 '22 at 15:50

0 Answers0