Plotting pie-plot with Pandas in Python

Asked Mar 18 '20 at 17:43

Active Mar 18 '20 at 17:45

Viewed 327 times

I started working with the Matplotlib pie chart and looks good so far! I am having only a few issue.

I am interested in a particular column and they are kind of categorical but numbers and these numbers mean different kinds of stuff that I would like to depict on the pie chart.

Here is my working code. I started off with a Python dictionary to map out the numbers

valuelables = {
    1:"Deutschland",
    2:"UK",
    3:"America"
}

s1 = df.request.value_counts(normalize=True)*100
size = s1.tolist()
keys = s1.keys().tolist()

In order for you to produce on your local computer:

print(size) # size gives a list [99.22871437721278, 0.5294334011693548, 0.2418522216178678]
print(keys) # keys give [2, 1, 3]

patches, texts = plt.pie(size, shadow=True, startangle=90, autopct='%1.1f%%')
plt.legend(patches, valuelables[keys], loc="best")
plt.axis('equal')

Unfortunately, I get the following error: TypeError: unhashable type: 'list'

This obviously comes from the dictionary mapping in the plt.legend(). How do I map it to produce a new pie plot like this below with the values on the plot like the second plot:

edited Mar 18 '20 at 17:45

asked Mar 18 '20 at 17:43

JA-pythonista

1,225
1
21
44

convert it to `tuple`. `tuple` is hashable since it is **immutable**. – Boseong Choi Mar 18 '20 at 17:45
`keys = tuple(s1.keys().tolist())` – Boseong Choi Mar 18 '20 at 17:45
convert what to tuple? – JA-pythonista Mar 18 '20 at 17:46
the `list`. `list` is unhashable because it is **mutable**. – Boseong Choi Mar 18 '20 at 17:47
now I get a new error: KeyError: (2, 1, 3) – JA-pythonista Mar 18 '20 at 17:49
1

The above is nonsense. What you're looking for is [mapping-dictionary-value-to-list](https://stackoverflow.com/questions/33078554/mapping-dictionary-value-to-list). Here, e.g. `plt.legend(patches, [valuelables[key] for key in keys])` – ImportanceOfBeingErnest Mar 18 '20 at 17:59
Note that `plt.pie` returns 3 lists when you fill in `autopct=`. So, `patches, texts, pct_texts = plt.pie(...)`. Optionally, you can also show the percentage in the legend: `plt.legend(patches, [valuelables[key] + ':' + p.get_text() for key, p in zip(keys, pct_texts)])` – JohanC Mar 18 '20 at 18:30
Thanks @ImportanceOfBeingErnest – JA-pythonista Mar 19 '20 at 07:29

Plotting pie-plot with Pandas in Python

0 Answers0