In [33]:
%matplotlib inline
import matplotlib.pyplot as plt
Two More Useful Python Libraries¶
Pandas¶
Pandas¶
- Python library for data manipulation and analysis
- Based on numpy, matplotlib
- Makes some common data analysis tasks very easy
- Documentation: pandas.pydata.org
In [34]:
import pandas as pd
pd._version.get_versions()
Out[34]:
In [35]:
import urllib.request
print(urllib.request.urlopen(
"http://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data").read().decode('ascii')[:100], '...')
In [36]:
df = pd.read_csv("http://archive.ics.uci.edu/ml/machine-learning-databases/iris/iris.data",
header=None,
names=['Sepal Length', 'Sepal Width', 'Petal Length', 'Petal Width', 'Species'])
In [37]:
type(df)
Out[37]:
Pandas DataFrames¶
- Row/Column datatype
- Essentially a matrix with labels and indexes
- Many useful operations
In [38]:
df.head(5)
Out[38]:
In [39]:
df.head(5).values ## extract numpy array
Out[39]:
In [40]:
df.head(5)['Petal Length'] ## extract entire columns by name
Out[40]:
In [41]:
df.head(5).Species
Out[41]:
In [42]:
df.head(5)[['Species', 'Petal Length']] ## extract more than one column
Out[42]:
In [43]:
df.T ## Transpose with labels
Out[43]:
In [44]:
df.index ## DataFrame rows are indexed
Out[44]:
In [45]:
df.ix[4] ## Extract individual rows by their index
Out[45]:
In [46]:
df.ix[[4, 17, 23]]
Out[46]:
In [47]:
df.head(5)
Out[47]:
In [48]:
df.head(5)['Sepal Length'] < 5 ## Boolean queries
Out[48]:
In [49]:
df[df['Sepal Length'] < 5].head(3) ## Boolean indexing
Out[49]:
In [50]:
df2 = df.copy() ## Cloning DataFrames
all_species = list(df.Species.unique())
print(all_species)
## Adding colums ## Transforming existing columns
df2['Type'] = df.Species.map(lambda s: all_species.index(s))
df2.head(5)
Out[50]:
In [51]:
df.describe() ## Quick summary statistics
Out[51]:
Note: 25th percentile = x $\rightarrow$ 25% of the data are $\leq$ x (50th percentile is the median)
In [52]:
df.groupby("Species").describe()
Out[52]:
In [53]:
df2.dot_color = df2.Species.map(lambda s: {'Iris-setosa': 'red',
'Iris-virginica': 'green',
'Iris-versicolor': 'blue'}[s])
## Quick plotting
df2.plot('Petal Width', 'Petal Length', kind='scatter', c=df2.dot_color);
"Box-and-whisker" plot¶
- Another way to examine feature distributions
- Reference: upload.wikimedia.org/.../Boxplot_vs_PDF.svg
In [54]:
df.plot(kind='box');
Seaborn: Statistical Data Visualization¶
- Produces advanced visualizations with much less work than using matplotlib directly
- Works together with pandas very well
- Documentation & tutorials: stanford.edu/~mwaskom/software/seaborn
In [55]:
import seaborn as sb
In [56]:
## Just put in a DataFrame object,
## seaborn does the right thing automatically:
sb.pairplot(df, hue='Species', size=3);
In [57]:
sb.lmplot(x='Sepal Width', y='Petal Length', data=df); # Does linear regression automatically
Plotting data facets on grids¶
In [61]:
g = sb.FacetGrid(df, col='Species', hue="Species")
g.map(plt.scatter, "Petal Length", "Petal Width");
In [59]:
## Add two boolean columns
df2 = df.copy()
df2['PL_low'] = df2['Petal Length'].map(lambda x: x < df2['Petal Length'].mean())
df2['SW_low'] = df2['Sepal Width'].map(lambda x: x < df2['Sepal Width'].mean())
df2.head(5)
Out[59]:
In [60]:
g = sb.FacetGrid(df2, col='PL_low', row='SW_low', hue='Species')
g.map(plt.scatter, "Petal Width", "Sepal Length");
