Make a histogram with a Gaussian curve in Python

Question

It’s the first time I do a Python histogram, but I’m not getting the same Python histogram I got in Excel. Also, I wanted to put a Gaussian curve and the mean with the uncertainty of it next to the graph as I saw in an article. The histogram I performed in Excel is as follows:

I’m trying to create a histogram where on the xx axis the size of the sides of the triangles I’m measuring appears in micrometers and on the yy axis the relative frequency at which these triangles appear. The data I used to obtain the Excel chart are as follows::

Tamanho  Frequência relativa
  0-1          0,282
  1-2          0,316
  2-3          0,171
  3-4          0,068
  4-5          0,085
  5-6          0,026
  6-7          0,026
  7-8          0,009
  8-9          0,009
  9-10         0,000
  10-11        0,000
  11-12        0,009

I would like to perform a Python histogram similar to this article ("Okada et al., Gas-Source CVD Growth of Atomic Layered WS2 from WF6 and H2S precursors with High Grain Size Uniformity, 2019"):

where the histogram, the Gaussian curve and the mean of the curve with the uncertainty appear. I tried to do it in Python, but I couldn’t get the relative frequency to appear on the yy axis of the histogram. I will put here my unsuccessful resolution attempt on Python made in a Jupyter Notebook of Anaconda:

import pandas as pd
from pathlib import Path
datafile=Path('/Users/maria/Desktop/graficosTese', 'amostra22Atriangulos.xlsx')
df=pd.read_excel(datafile, 
                 index_col=0)
df.head()
df.hist(column='Frequência relativa')

My Excel file has the following structure:

Could someone help me do the histogram with the Gaussian curve where the average equation with the uncertainty in the graph was? Thank you very much

Error obtained now:

Answer 1

Since you have the frequency data ready, I believe that you do not need to make use of the functions that create histograms, such as plt.hist of matplotlib and sns.distplot of Seaborn. Anyway, it is worth a read in the documentation since this is the most common way to build histograms (matplotlib, Seaborn).

In case of your problem, however, just a bar chart:

import pandas as pd
import matplotlib.pyplot as plt
from unidecode import unidecode
import numpy as np

df=pd.DataFrame({'frequencia_relativa': {'0-1': 0.282, '1-2': 0.316, '2-3': 0.171, '3-4': 0.068, '4-5': 0.085, '5-6': 0.026, '6-7': 0.026, '7-8': 0.009, '8-9': 0.009, '9-10': 0.0, '10-11': 0.0, '11-12': 0.009}})

df=df.reindex(['0-1', '1-2', '2-3', '3-4', '4-5', '5-6', '6-7',
       '7-8', '8-9', '9-10', '10-11', '11-12'])


fig, ax=plt.subplots()
ax.bar(x=df.index, height=df.frequencia_relativa)
ax.set_yticklabels(["{:.2f}".format(k).replace(".",",") for k in np.arange(0.,0.35,0.05)])
plt.show()

The second part of the problem (i.e., plotting a Gaussian distribution) is also possible to do manually, although my indication is that you use a kernel. Following example (using the average and standard deviation parameters of the spreadsheet):

import pandas as pd
import matplotlib.pyplot as plt
import numpy as np
from scipy.stats import norm

df=pd.DataFrame({'frequencia_relativa': {'0-1': 0.282, '1-2': 0.316, '2-3': 0.171, '3-4': 0.068, '4-5': 0.085, '5-6': 0.026, '6-7': 0.026, '7-8': 0.009, '8-9': 0.009, '9-10': 0.0, '10-11': 0.0, '11-12': 0.009}})
df=df.reindex(['0-1', '1-2', '2-3', '3-4', '4-5', '5-6', '6-7',
       '7-8', '8-9', '9-10', '10-11', '11-12'])

fig, ax=plt.subplots()
ax.bar(x=df.index, height=df.frequencia_relativa, alpha=0.5)
ax.set_yticklabels(["{:.2f}".format(k).replace(".",",") for k in np.arange(0.,0.35,0.05)])

x_axis = np.arange(0, 10, 0.001)
#normal com média 2,232 e variância 1.88, conforme planilha
plt.plot(x_axis, norm.pdf(x_axis,2.232,1.88))
plt.show()