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

# # Data Visualization with Python
# 
# ## Part 1: Python + Matplotlib
# 
# ### [Guy Allard](mailto://w.g.allard@lumc.nl)

# # Matplotlib
# - Plotting library for Python
# - High quality figures suitable for publication
# - Integrates with IPython, Jupyter and NumPy (in PyLab mode)
# - Established and robust
# - Large community / user base

# # Interfaces
# 1. Object Oriented  
#   - Best for larger development projects
#   - Have to keep track of figures and axes
#   - Steep learning curve
# <br><br>
# 2. Pyplot State Machine  
#   - For interactive plotting
#   - Takes care of many housekeeping tasks
#   - Easier to learn than the OO interface
# <br><br>
# 3. Pylab
#   - Modelled on matlab
#   - Imports common modules
#   - Handles most housekeeping tasks
#   - Easiest to learn
#   - The one we will be using!

# # Interfaces Example
# 
# 1. Object-oriented interface
# ```python
# import matplotlib.pyplot as plt
# import numpy as np
# x = np.arange(0, 10, 0.2)
# y = np.sin(x)
# fig = plt.figure()
# ax = fig.add_subplot(111)
# ax.plot(x, y)
# ```
# 
# 2. State-machine environment (pyplot)
# ```python
# import matplotlib.pyplot as plt
# import numpy as np
# x = np.arange(0, 10, 0.2)
# y = np.sin(x)
# plt.plot(x, y)
# ```
# 
# 3. PyLab mode
# ```python
# %pylab
# x = arange(0, 10, 0.2)
# y = sin(x)
# plot(x, y)
# ```

# # Getting help
# 
# Consult the built-in documentation, for example:
# ```
# >>> help(subplot)
# Help on function subplot in module matplotlib.pyplot:
# 
# subplot(*args, **kwargs)
#     Return a subplot axes positioned by the given grid definition.
# ...
# ```
# 
# 

# # Useful Resources
# - Matplotlib Homepage
#   - https://matplotlib.org/
# <br><br>
# - Gallery
#   - https://matplotlib.org/gallery.html
#   - Many examples with source code
# <br><br>
# - Online documentation
#   - https://matplotlib.org/contents.html
#   - Full API documentation

# # First Steps
# 
# ## Preparing the Jupyter Notebook
# 1. Open a new Jupyter Notebook
# 2. Run this code in the first empty cell:
# ```
# %pylab inline
# ```
# 3. Now any pylab plotting commands will display in the notebook

# In[1]:


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


# # Grab some data
# Use Pandas to load a dataset which contains population data for four countries

# In[2]:


import pandas as pd

populations = pd.read_csv(
    'https://git.lumc.nl/courses/programming-course/raw/visualization-2018/visualization/data/populations.csv'
)


# Take a quick look at the data

# In[3]:


populations.head()


# # Plot it!
# Let's make a plot the population of the Netherlands on the y-axis, and the year on the x-axis

# In[4]:


plot(populations['Year'], populations['Netherlands']);


# # Add titles and label the axes

# In[5]:


plot(populations['Year'], populations['Netherlands'])

title('Historical Population of The Netherlands')

xlabel('Year')

ylabel('Population (Millions)');


# # Change some properties of the line
# 
# How about a 5px thick orange line?

# In[6]:


plot(populations['Year'], populations['Netherlands'], 
     linewidth=5, color='orange')

title('Historical Population of The Netherlands')

xlabel('Year')

ylabel('Population (Millions)');


# # Change some properties of the x-axis
# 
# Label at five-year intervals  
# Display the label vertically

# In[7]:


plot(populations['Year'], populations['Netherlands'], 
     linewidth=5, color='orange')

title('Historical Population of The Netherlands')
xlabel('Year')
ylabel('Population (Millions)')

xticks(range(1950, 2016, 5), rotation=90);


# # Change which years are displayed

# In[8]:


plot(populations['Year'], populations['Netherlands'], 
     linewidth=5, color='orange')

title('Historical Population of The Netherlands')
xlabel('Year')
ylabel('Population (Millions)')
xticks(range(1950, 2016, 5), rotation=90)

xlim(1970, 1990);


# # Change the y-axis scale

# In[9]:


plot(populations['Year'], populations['Netherlands'], 
     linewidth=5, color='orange')

title('Historical Population of The Netherlands')
xlabel('Year')
ylabel('Population (Millions)')
xticks(range(1950, 2016, 5), rotation=90)
xlim(1970, 1990)

ylim(13,15);


# # Clean up the number formatting on the y-axis
# 
# Integer tick labels

# In[10]:


plot(populations['Year'], populations['Netherlands'], 
     linewidth=5, color='orange')

title('Historical Population of The Netherlands')
xlabel('Year')
ylabel('Population (Millions)')
xticks(range(1950, 2016, 5), rotation=90)
xlim(1970, 1990)
ylim(13,15)

yticks(range(13,16));


# # Plot multiple series
# 
# Calling **plot** multiple times within the same cell will add multiple series to the chart
# 
# Let's compare the Dutch with the Danes

# In[11]:


plot(populations['Year'], populations['Netherlands'], color='orange')
plot(populations['Year'], populations['Denmark'], color='red')

title('Historical Populations of The Netherlands and Denmark')
xlabel('Year')
ylabel('Population (Millions)')
xticks(range(1950, 2016, 5), rotation=90);


# # Add a legend
# 
# 1. Give each plotted line a label
# 2. Add a legend to the figure

# In[12]:


plot(populations['Year'], populations['Netherlands'],
     color='orange', label='The Netherlands')

plot(populations['Year'], populations['Denmark'],
     color='red', label='Denmark')

legend(loc='upper left')

title('Historical Populations of The Netherlands and Denmark')
xlabel('Year')
ylabel('Population (Millions)')
xticks(range(1950, 2016, 5), rotation=90);


# # Other plot types
# 
# Let's load a different dataset and take a look at some different plot types

# In[13]:


flowers = pd.read_csv('https://git.lumc.nl/courses/programming-course/raw/visualization-2018/visualization/data/iris.csv')

flowers.head()


# # Boxplots
# 
# A simple boxplot of the sepal-length distribution

# In[14]:


boxplot(flowers['sepal_length'], labels=['Sepal_length']);


# # Boxplots
# 
# Distributions of multiple features

# In[15]:


# make a list containing the numeric feature column names
features = list(flowers.columns[:-1])
features


# In[16]:


# plot the data
boxplot([flowers[f] for f in features], labels=features);


# # Controlling the size of the plot
# 
# Let's change the shape of the boxplot

# In[17]:


# make the figure 10 'units' wide and 5 'units' high
figsize(10, 5)

# plot the data
boxplot([flowers[f] for f in features], labels=features);


# In[18]:


figsize(7,4)


# # Histogram

# In[19]:


hist(flowers['petal_length'])

title('Petal Length Distribution')
xlabel('petal length')
ylabel('count');


# # Histogram
# 
# change the number of 'bins'

# In[20]:


hist(flowers['petal_length'], bins=20)

title('Petal Length Distribution')
xlabel('petal length')
ylabel('count');


# # Histogram
# 
# Some formatting
# 
# - Lines around the bars
# - color

# In[21]:


hist(flowers['petal_length'], bins=20, facecolor='teal', edgecolor='black', alpha=0.7)

title('Petal Length Distribution')
xlabel('petal length')
ylabel('count');


# # Subplots
# 
# Separate plots with their own axes within a single figure
# 
# The syntax can be confusing!

# In[22]:


for i in range(1, 5):
    subplot(2, 2, i)
    xticks([]), yticks([])
    text(0.5, 0.5, 'subplot(2, 2, %d)' % i, ha='center', size=18, alpha=0.75);


# subplot(2, 2, 1) indicates the first cell of a 2 row x 2 column matrix
# 
# subplot(2, 2, 4) indicates the fourth cell of a 2 column x 2 row matrix

# # Subplots
# 
# More complicated layouts

# In[23]:


subplot(1, 3, 1)            # 1 row, 3 columns, cell 1
xticks([]), yticks([])
text(0.5, 0.5, '(1, 3, 1)', ha='center', size=18, alpha=0.75)

subplot(2, 3, 3)            # 2 rows, 3 columns, cell 3
xticks([]), yticks([])
text(0.5, 0.5, '(2, 3, 3)', ha='center', size=18, alpha=0.75)

subplot(3, 2, 6)            # 3 rows, 2 columns, cell 6
xticks([]), yticks([])
text(0.5, 0.5, '(3, 2, 6)', ha='center', size=18, alpha=0.75)

subplot(3, 3, 5)            # 3 rows, 3 columns, cell 5
xticks([]), yticks([])
text(0.5, 0.5, '(3, 3, 5)', ha='center', size=18, alpha=0.75);


# # Subplots and Boxplots
# 
# Compare how the features are distributed by species

# In[24]:


species = list(set(flowers.species))
print(species)


# In[25]:


# make a dataset for each species
setosa = flowers[flowers.species == 'setosa']
versicolor = flowers[flowers.species == 'versicolor']
virginica = flowers[flowers.species == 'virginica']


# In[26]:


figsize(10, 8)
for cell, feature in enumerate(features):
    subplot(2, 2, cell + 1)
    boxplot(
        [setosa[feature], versicolor[feature], virginica[feature]],
        labels=species
    )
    ylabel(feature)


# In[27]:


figsize(7,4)


# # Sketch-style drawing
# 
# using xkcd mode

# In[28]:


with xkcd():
    hist(flowers['petal_length'], bins=20, facecolor='teal', edgecolor='black')
    title('Petal Length Distribution')
    xlabel('petal length')
    ylabel('count');


# # Saving to a file
# 
# Images can be saved to a file using savefig after the plotting commands:
# ```
# savefig('myplot.pdf')
# ```
# 
# The format of the saved image will be inferred from the given file extension.

# # The End
# 
# This lesson was based on previous work by [Jeroen Laros](mailto://j.f.j.laros@lumc.nl) and Martijn Vermaat
# 
# License: [Creative Commons Attribution 3.0 License (CC-by)](http://creativecommons.org/licenses/by/3.0)

# In[ ]: