Visualization Cheat Sheet.py

import seaborn as sns
import pandas as pd
import matplotlib.pyplot as plt
%matplotlib inline

#adjust seaborn plot size
plt.figure(figsize=(20,12))

#pairplot (my favorite) - Draw scatterplots for joint relationships and histograms for univariate distributions
#hue optional like on all other seaborn plots
sns.pairplot(MyDataFrame,hue='some_categorical_column')

"""
barplot for aggregation of categorical variables. - the estimator argument is what determines the aggregation function(done with numpy)
"""
sns.barplot(x='catgegorical_column_name',y='column_name_to_aggregate',data=MyDataFrame,estimator=np.sum) #np.sum should sum the y for x

#Count plot - count a categorical variable
sns.countplot(x='categorical_column_to_count',data=MyDataFrame)

"""
stripplot - Draw a scatterplot where one variable is categorical
Categorical variables go on the x axis. Continuous variable on the y axis.
jitter argument staggers the points so you can see them better
hue argument takes a categorical variable to add another level of detail
"""
sns.stripplot(x='sex',y='goals_scored',data=MyDataFrame,jitter=True,hue='occupation')

"""
distplot - Flexibly plot a univariate distribution of observations.
"""
sns.distplot(df['column_name'],bins=30)

#jointplot
sns.jointplot(x='continuous_column1',y='continuous_column2',data=MyDataFrame,kind='scatter')

#heatmaps
#this is nifty in conjunction with the Pandas .corr() method for a heatmap of Pearson Correlation Coefficients
sns.heatmap(MyDataFrame)

#regplot - regression plot for 2 variables
sns.regplot(x='column1',y='column2',data=MyDataFrame)

#Scatterplot - built in to pandas
#c argument is optional. You can put a column name with discrete values as an argument to add more details
df.plot.scatter(x='blah',y='blahblah',c='discrete_value')

#line plot
df.plot.line(x='blah',y='blahblah')

#histogram - I like this one, numeric data only
#bin=10 is a default argument
df['column'].plot.hist()