#!/usr/bin/env python
# coding: utf-8

# 
# <table>
# <tr>
# <td width=15%><img src="./img/UGA.png"></img></td>
# <td><center><h1>Introduction to Python for Data Sciences</h1></center></td>
# <td width=15%><a href="http://www.iutzeler.org" style="font-size: 16px; font-weight: bold">Franck Iutzeler</a> </td>
# </tr>
# </table>
# 
# 

# 
# <br/><br/>
# 
# <center><a style="font-size: 40pt; font-weight: bold">Chap. 3 - Data Handling with Pandas </a></center> 
# 
# <br/><br/>
# 

# 
# 
# # 3- Fancy Visualization with Seaborn
# 

# ## Advanced visualization
# 
# 
# [Seaborn](https://seaborn.pydata.org/) is a package that produces somewhat nicer and more data oriented plots than Matplotlib. It also gives a fresher look to matlotlib plots.

# In[1]:


import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

get_ipython().run_line_magic('matplotlib', 'inline')


# In[2]:


# Create some data
rng = np.random.RandomState(0)
x = np.linspace(0, 10, 500)
y = np.cumsum(rng.randn(500, 3), 0)


# In[3]:


plt.plot(x, y)
plt.legend('one two three'.split(' '));


# Let us import seaborn and change the matplotlib style with <tt>sns.set()</tt>

# In[4]:


import seaborn as sns
sns.set()


# In[5]:


# Same command but now seaborn is set
plt.plot(x, y)
plt.legend('one two three'.split(' '));


# ### Plotting Distributions
# 
# Apart from the standard histograms <tt>plt.hist</tt>, Seaborn provides smoothed density plots based on data using <tt>sns.kdeplot</tt> or <tt>sns.displot</tt>. 

# In[6]:


data = np.random.multivariate_normal([0, 1.5], [[1, 0.2], [0.2, 2]], size=2000)
data = pd.DataFrame(data, columns=['x', 'y'])

for col in 'xy':
    plt.hist(data[col], alpha=0.5) # alpha=0.5 provides semi-transparent plots


# <tt>kdeplot</tt> provides density plots from an array or series (<tt>shade=True</tt> provide filled ones).

# In[15]:


sns.kdeplot(data['x'])
sns.kdeplot(data['y'],shade=True)


# <tt>displot</tt> is a mix of the two previous ones.

# In[16]:


sns.displot(data['x'])
sns.histplot(data['y'])


# Two-dimensional dataset may be represented by level sets with <tt>kdeplot</tt>.

# In[19]:


sns.kdeplot(data['x'],y = data['y'], shade=True, thresh=0.05, cmap="Reds", cbar=True)


# Joint distribution and the marginal distributions can be displayed together using <tt>jointplot</tt>

# In[21]:


sns.jointplot(x= "x", y= "y", data = data, kind='kde');


# ### Exploring features correlations and interest to classification
# 
# Seaborn provides an efficient tool for quickly exploring different features and classification with <tt>pairplot</tt>.

# In[22]:


import pandas as pd
import numpy as np

iris = pd.read_csv('data/iris.csv')
print(iris.shape)
iris.head()


# In[23]:


sns.pairplot(iris, hue='species')


# <tt>factorplot</tt> also provides error plots.

# In[25]:


sns.catplot( x = "species" ,  y="sepal_length"  , data=iris ,   kind="box")


# ### Melting dataframes
# 
# For displaying classification data, it is sometimes interesting to **melt** dataframes, that is separating 
# * **id:** the classes typically, things that are not numeric, that have to be kept in place (in our case with *iris*, the species)
# * **values:** the *columns* corresponding to values (in our case with *iris*, the sepal_length, sepal_width, etc.)
# 
# The command <tt>pd.melt</tt> return a dataframe with as columns: the id, the variable (former column) name, and associated value.

# In[26]:


irisS = pd.melt(iris,id_vars="species",value_vars=["sepal_length","sepal_width","petal_length","petal_width"])
irisS.head()


# In[28]:


sns.catplot( x=  "species" , y = "value" , col="variable"  , data=irisS , kind="box")