Data Visualization

Structured data

Different kinds of plots; the bottom row uses a different dataset than the other rows

Visualizing data can help us get insights into it. Depending on the type of data you want to visualize, you may do so in different ways:

For 1D data, like a temperature measurement or the number of hits on a webpage we use line plots
For 2D data, like the weight and height of a person we use scatter plots
To see the distribution of your data you can use histograms
To see how much of our data falls into a certain category we can use cumulative plots or pie plots
To see the spread of our data we can use box plots
To see the estimated kernel(probability) density of our variables we can use kernel density estimation plots
To represent the relationship between 2 numerical variables we can use a Hexbin plot

Unstructured data

Examples: images, sound, text

On the left we have a spectogram of the sound, created using a Fast Fourier Transform.

Image data visualization with Matplotlib

plt.figure(figsize=(16, 16))

from random import randint

rnd = randint(0, len(X_orig)-16)

j = 0
for i in range(rnd, rnd+16):
    img = np.array(X_orig.iloc[i])
    img = img.reshape(28,28)

    plt.subplot(4, 4, j+1)
    j=j+1
    s = 'L: ' +str(y_orig.iloc[i])
    plt.gca().set_title(s)
    plt.imshow(img)

plt.tight_layout()