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

# ##Cufflinks
# 
# This library binds the power of [plotly](http://www.plot.ly) with the flexibility of [pandas](http://pandas.pydata.org/) for easy plotting.
# 
# This library is available on https://github.com/santosjorge/cufflinks
# 
# This tutorial assumes that the plotly user credentials have already been configured as stated on the [getting started](https://plot.ly/python/getting-started/) guide.

# In[1]:


import pandas as pd
import cufflinks as cf
import numpy as np


# In[2]:


get_ipython().run_line_magic('reload_ext', 'autoreload')
get_ipython().run_line_magic('autoreload', '2')


# We make all charts public and set a global theme

# In[3]:


cf.set_config_file(world_readable=True,theme='pearl')


# We create a set of timeseries

# In[26]:


df=pd.DataFrame(np.random.randn(100,5),index=pd.date_range('1/1/15',periods=100),
                columns=['IBM','MSFT','GOOG','VERZ','APPL'])
df=df.cumsum()


# **iplot** can be used on any DataFrame to plot on a plotly chart. 
# If no filename is specified then a generic *Plotly Playground* file is created.
# 
# All the charts are created as private by default. To make them public you can use **world_readable=True**
# 
# Let's look at the avilable parameters

# In[5]:


help(df.iplot)


# In[27]:


df.iplot(filename='Tutorial 1')


# ### Customizing Themes
# 
# We can pass a **theme** to the **iplot** function. 
# 3 themes are available, but you can create your own
# * Solar
# * Pearl (Default)
# * White

# In[28]:


df[['APPL','IBM','VERZ']].iplot(theme='white',filename='Tutorial White')


# We can also pass common metadata for the chart

# In[29]:


df.iplot(theme='pearl',filename='Tutorial Metadata',title='Stock Returns',xTitle='Dates',yTitle='Returns')


# ### Bestfit Lines
# 
# We can easily add a bestfit line to any Series
# 
# This will automatically add a best fit approximation and the equation as the legend.

# In[30]:


df['IBM'].iplot(filename='IBM Returns',bestfit=True)


# ### Customizing Colors
# 
# We can pass any color (either by Hex, RGB or Text *)  
# 
# *Text values are specified in the cufflinks.colors modules

# In[31]:


df['IBM'].iplot(filename='IBM Returns - colors',bestfit=True,colors=['pink'],bestfit_colors=['blue'])


# ### Filled Traces
# 
# We can add a fill to a trace with **fill=True**

# In[32]:


df['IBM'].iplot(filename='Tutorial Microsoft',fill=True,colors=['green'])


# ### Bar Charts
# 
# We can easily create a bar chart with the parameter **kind**

# In[33]:


df.sum().iplot(kind='bar',filename='Tutorial Barchart')


# Bars can also be stacked by a given dimension

# In[34]:


df.resample('M').iplot(kind='bar',barmode='stacked',filename='Tutorial Bar Stacked')


# ### Spread and Ratio charts
# 
# We can also create spread and ratio charts on the fly with **kind='spread'** and **kind='ratio'**

# In[35]:


df[['VERZ','IBM']].iplot(filename='Tutorial Spread',kind='spread')


# In[55]:


(df[['GOOG','MSFT']]+20).iplot(filename='Tutorial Ratio',kind='ratio',colors=['green','red'])


# ### Annotations
# 
# Annotations can be added to the chart and these are automatically positioned correctly. 
# 
# **Annotations** should be specified in a dictionary form

# In[56]:


annotations={'2015-01-15':'Dividends','2015-03-31':'Split Announced'}
df['MSFT'].iplot(filename='Tutorial Annotations',annotations=annotations)


# ### Output as Image
# 
# The output of a chart can be in an image mode as well. 
# 
# For this we can use **asImage=True**
# 
# We can also set the dimensions (optional) with **dimensions=(width,height)**

# In[61]:


df[['VERZ','MSFT']].iplot(filename='Tutorial Image',theme='white',colors=['pink','blue'],asImage=True,dimensions=(800,500))


# ### Advanced Use
# 
# It is also possible to get the Plotly Figure as an output to tweak it manually
# 
# We can achieve this with **asFigure=True**

# In[62]:


df['GOOG'].iplot(asFigure=True)


# We can also get the **Data** object directly

# In[63]:


data=df.to_iplot()


# In[64]:


data[0]['name']='My Custom Name'


# And pass this directly to **iplot**

# In[66]:


df.iplot(data=data,filename='Tutorial Custom Name')


# In[ ]: