Question

Histograms, Box Plots, and Violin Plots Q06.50 Plot the histogram of a normal distribution of 100 random numbers. Use NumPy's np.random.normal() function to create the array of numbers. Set a mean u = 20 and a standard deviation o = 7.

pls use codes that run in Jupyter notebook

Answer 1

You can use Matplotlib library's, hist() function from pyplot module to draw a histogram easily. If you don't have it installed, simply write the following line in the command line: pip install matplotlib.

So here's the small code, you need to execute inside the jupyter notebook:

Code:

<start of code>

import matplotlib.pyplot as plt

import numpy as np

%matplotlib inline

mu = 20
sigma = 7

numbers = np.random.normal(mu, sigma, size=100)

plt.hist(numbers, 20)

plt.show()

<end of code>

That's all you need to write to get the histogram. So let me just brief you about the function and the arguments expected by those functions.

1. The alias plt is commonly used for matplotlib's pyplot library.
2. Similarly the alias np is used for numpy library.
3. If using a Jupyter notebook, you need to include the line %matplotlib inline below the imports.
4. Then we'll use numpy's np.random.normal() function to produce an array of random numbers with a normal distribution. The general format of the np.random.normal() function is:

   var = np.random.normal(mean, standard_deviation, size=<number of values>)

5. Matplotlib's plt.hist() function produces histogram plots. The first positional argument passed to plt.hist() is a list or array of values, the second positional argument denotes the number of bins on the histogram. You can also pass additional arguments like denisty, histtype, alpha (opacity), etc.

AND HERE'S OUR OUTPUT:

" 5 10 15 20 25 30 35

IN CASE YOU WANT TO CHANGE THE TYPE OF HISTOGRAM, YOU CAN USE THE "histtype" PARAMETER LIKE: plt.hist(numbers, 20, histype='step'). Other types of histograms are - 'barstacked', 'stepfilled'.

IN THIS CASE THE OUTPUT WOULD BE LIKE:

Also, you can draw histograms using other libraries too like seaborn, numpy, etc. I myself prefer matplotlib. You can try out other ones too. But they are altogether similar. For instance, seaborn library actually uses matplotlib to draw plots.

Hope this helps. Best of luck!!