Can someone please help me make my overlapping histograms?

Question

I have three dataframes in the format below, which have a column with the month of the year in digit format, and a column adjacent to it which has the number of items occurring in that month. I wanted to create an overlapping histogram detailing the spread between the three histograms but for some reason I keep getting the same thing!

    month_box   Sum Value
0     1        4812
1     2        2053
2     3        2405
3     4        2353
4     5        2427
5     6        2484
6     8        2579
7     9        2580
8    10        2497
9    11        2510
10   12        2202

The code I am using is below:

sns.distplot(bex_boxdf['month_box'],kde=False,label = 'Bexley')
sns.distplot(west_boxdf['month_box'],kde=False,label = 'Westminster')
sns.distplot(gwch_boxdf['month_box'],kde=False,label = 'Greenwich')
plt.legend(prop={'size': 12})
plt.title('Crime by month')
plt.xlabel('Month')
plt.ylabel('Density')

I attach below the result I get...help would be appreciated thank you.

Answer 1

Using the data provided by @Esa, here are three different views using Matplotlib.

There is also a 'stepfilled' histogram type that I didn't include but could be useful depending on the distribution of the data:

import pandas as pd
import matplotlib.pyplot as plt

df = pd.DataFrame({'month_box': [1,2,3,4,5,6,7,8,9,10,11,12],
               'Bexley_sum': [4812,2053,2405,2353,2427,2484,2579,2580,
                             2497,2510,2202,2021],
               'Westminster_sum': [4712,2050,2435,2323,2487,2414,2679,2780,
                             2490,2110,2702,2022],
               'Greenwich_sum': [4812,2053,2405,2353,2427,2484,2579,2580,
                             2497,2510,2202,2021],
               })

data = df["Bexley_sum"], df["Westminster_sum"], df["Greenwich_sum"]
labels = ["Bexley", "Westminster", "Greenwich"]
fig, ax = plt.subplots(ncols=3, sharex=True, sharey=True)
ax[0].hist(x=data, histtype="bar",label=labels)
ax[1].hist(x=data, histtype="barstacked",label=labels)
ax[2].hist(x=data, histtype="step", label=labels)
plt.legend()
plt.show()

Matplotlib has a lot of customization options. Referring to the documentation could be useful.

https://matplotlib.org/stable/api/_as_gen/matplotlib.pyplot.hist.html

Answer 2

Sometimes it is easier to plot directly from the dataframe or using matplotlib instead of seaborn. And other times seaborn is better, so its best to try to learn both at some level at least.

Here's a simple solution if you first arrange your data into one dataframe. You did not provide the other two dataframes so I set some values as an example.

df = pd.DataFrame({'month_box': [1,2,3,4,5,6,7,8,9,10,11,12], 
               'Bexley_sum': [4812,2053,2405,2353,2427,2484,2579,2580,
                             2497,2510,2202,2021],
               'Westminster_sum': [4712,2050,2435,2323,2487,2414,2679,2780,
                             2490,2110,2702,2022],                   
               'Greenwich_sum': [4812,2053,2405,2353,2427,2484,2579,2580,
                             2497,2510,2202,2021],                   
               })

df.plot(x='month_box', y=['Bexley_sum', 'Westminster_sum', 'Greenwich_sum'], kind='bar')