Normalization in Python/v3

Learn how to normalize data by fitting to intervals on the real line and dividing by a constant

Note: this page is part of the documentation for version 3 of Plotly.py, which is not the most recent version.
See our Version 4 Migration Guide for information about how to upgrade.

New to Plotly?¶

Plotly's Python library is free and open source! Get started by dowloading the client and reading the primer.
You can set up Plotly to work in online or offline mode, or in jupyter notebooks.
We also have a quick-reference cheatsheet (new!) to help you get started!

Imports¶

The tutorial below imports NumPy, Pandas, and SciPy.

In [1]:

import plotly.plotly as py
import plotly.graph_objs as go
import plotly.tools as tools
from plotly.tools import FigureFactory as FF

import numpy as np
import pandas as pd
import scipy

Import Data¶

To properly visualize our data and normalization, let us import a dataset of Apple Stock prices in 2014:

In [2]:

apple_data = pd.read_csv('https://raw.githubusercontent.com/plotly/datasets/master/2014_apple_stock.csv')
df = apple_data[0:10]

table = FF.create_table(df)
py.iplot(table, filename='apple-data-sample')

Out[2]:

Normalize by a Constant¶

Normalize a dataset by dividing each data point by a constant, such as the standard deviation of the data.

In [4]:

data = apple_data['AAPL_y']

data_norm_by_std = [number/scipy.std(data) for number in data]

trace1 = go.Histogram(
    x=data,
    opacity=0.75,
    name='data'
)

trace2 = go.Histogram(
    x=data_norm_by_std,
    opacity=0.75,
    name='normalized by std = ' + str(scipy.std(data)),
)

fig = tools.make_subplots(rows=2, cols=1)

fig.append_trace(trace1, 1, 1)
fig.append_trace(trace2, 2, 1)

fig['layout'].update(height=600, width=800, title='Normalize by a Constant')
py.iplot(fig, filename='apple-data-normalize-constant')

This is the format of your plot grid:
[ (1,1) x1,y1 ]
[ (2,1) x2,y2 ]

Out[4]:

Normalize to [0, 1]¶

Normalize a dataset by dividing each data point by the norm of the dataset.

In [5]:

data_norm_to_0_1 = [number/scipy.linalg.norm(data) for number in data]

trace1 = go.Histogram(
    x=data,
    opacity=0.75,
    name='data',
)

trace2 = go.Histogram(
    x=data_norm_to_0_1,
    opacity=0.75,
    name='normalized to [0,1]',
)

fig = tools.make_subplots(rows=2, cols=1)

fig.append_trace(trace1, 1, 1)
fig.append_trace(trace2, 2, 1)

fig['layout'].update(height=600, width=800, title='Normalize to [0,1]')
py.iplot(fig, filename='apple-data-normalize-0-1')

This is the format of your plot grid:
[ (1,1) x1,y1 ]
[ (2,1) x2,y2 ]

Out[5]:

Normalizing to any Interval¶

Normalize a dataset to an interval [a, b] where a, b are real numbers.

In [6]:

a = 10
b = 50
data_norm_to_a_b = [(number - a)/(b - a) for number in data]

trace1 = go.Histogram(
    x=data,
    opacity=0.75,
    name='data',
)

trace2 = go.Histogram(
    x=data_norm_to_a_b,
    opacity=0.75,
    name='normalized to [10,50]',
)

fig = tools.make_subplots(rows=2, cols=1)

fig.append_trace(trace1, 1, 1)
fig.append_trace(trace2, 2, 1)

fig['layout'].update(height=600, width=800, title='Normalize to [10,50]')
py.iplot(fig, filename='apple-data-normalize-a-b')

This is the format of your plot grid:
[ (1,1) x1,y1 ]
[ (2,1) x2,y2 ]

Out[6]: