In [3]:
# Data Source: https://www.kaggle.com/worldbank/world-development-indicators
# Folder: 'world-development-indicators'
In [1]:
import pandas as pd
import numpy as np
import random
import matplotlib.pyplot as plt
In [3]:
data = pd.read_csv('./world-development-indicators/Indicators.csv')
data.shape
Out[3]:
In [5]:
countries = data['CountryName'].unique().tolist()
indicators = data['IndicatorName'].unique().tolist()
This is a really large dataset, at least in terms of the number of rows. But with 6 columns, what does this hold?
In [6]:
data.head(2)
Out[6]:
Looks like it has different indicators for different countries with the year and value of the indicator.
We already saw how the USA's per-capita CO2 production related to other countries, let's see if we can find some more indicators in common between countries.
To have some fun, we've picked countries randomly but then stored our random results so you can rerun it with the same answers.
In [7]:
# Filter 1
# Picks years of choice
yearsFilter = [2010, 2011, 2012, 2013, 2014]
In [8]:
# Filter 2
# Pick 2 countries randomly
countryFilter = random.sample(countries, 2)
In [9]:
countryFilter
Out[9]:
In [10]:
# Filter 3
# Pick 1 Indicator randomly
indicatorsFilter = random.sample(indicators, 1)
indicatorsFilter
Out[10]:
In [11]:
filterMesh = (data['CountryName'] == countryFilter[0]) & (data['IndicatorName'].isin(indicatorsFilter)) & (data['Year'].isin(yearsFilter))
country1_data = data.loc[filterMesh]
In [12]:
len(country1_data)
Out[12]:
In [13]:
filterMesh = (data['CountryName'] == countryFilter[1]) & (data['IndicatorName'].isin(indicatorsFilter)) & (data['Year'].isin(yearsFilter))
country2_data = data.loc[filterMesh]
In [15]:
len(country2_data)
Out[15]:
The code below will randomly pick countries and indicators until it finds two countries who have data for an indicator over this time frame. We used it to produce the fixed values you see later, feel free to play with this yourself!
In [17]:
filteredData1 = []
filteredData2 = []
In [18]:
'''
Plot:
countryFilter: pick two countries,
indicatorsFilter: pick an indicator,
yearsFilter: plot for years in yearsFilter
'''
# problem - not all countries have all indicators so if you go to visualize, it'll have missing data.
# randomly picking two indicators and countries, do these countries have valid data over those years.
# brings up the discussion of missing data/ missing fields
# until we find full data
while(len(filteredData1) < len(yearsFilter)-1):
# pick new indicator
indicatorsFilter = random.sample(indicators, 1)
countryFilter = random.sample(countries, 2)
# how many rows are there that have this country name, this indicator, and this year. Mesh gives bool vector
filterMesh = (data['CountryName'] == countryFilter[0]) & (data['IndicatorName'].isin(indicatorsFilter)) & (data['Year'].isin(yearsFilter))
# which rows have this condition to be true?
filteredData1 = data.loc[filterMesh]
filteredData1 = filteredData1[['CountryName','IndicatorName','Year','Value']]
# need to print this only when our while condition is true
if(len(filteredData1) < len(yearsFilter)-1):
print('Skipping ... %s since very few rows (%d) found' % (indicatorsFilter, len(filteredData1)))
In [19]:
# What did we pick eventually ?
indicatorsFilter
Out[19]:
In [20]:
len(filteredData1)
Out[20]:
In [21]:
'''
Country 2
'''
while(len(filteredData2) < len(filteredData1)-1):
filterMesh = (data['CountryName'] == countryFilter[1]) & (data['IndicatorName'].isin(indicatorsFilter)) & (data['Year'].isin(yearsFilter))
filteredData2 = data.loc[filterMesh]
filteredData2 = filteredData2[['CountryName','IndicatorName','Year','Value']]
#pick new indicator
old = countryFilter[1]
countryFilter[1] = random.sample(countries, 1)[0]
if(len(filteredData2) < len(filteredData1)-1):
print('Skipping ... %s, since very few rows (%d) found' % (old, len(filteredData2)))
In [22]:
if len(filteredData1) < len(filteredData2):
small = len(filteredData1)
else:
small = len(filteredData2)
In [23]:
filteredData1=filteredData1[0:small]
filteredData2=filteredData2[0:small]
In [24]:
filteredData1
Out[24]:
In [25]:
filteredData2
Out[25]:
Matplotlib: Additional Examples
In [26]:
%matplotlib inline
import matplotlib.pyplot as plt
fig, axis = plt.subplots()
# Grid lines, Xticks, Xlabel, Ylabel
axis.yaxis.grid(True)
axis.set_title(indicatorsFilter[0],fontsize=10)
axis.set_xlabel(filteredData1['CountryName'].iloc[0],fontsize=10)
axis.set_ylabel(filteredData2['CountryName'].iloc[0],fontsize=10)
X = filteredData1['Value']
Y = filteredData2['Value']
axis.scatter(X, Y)
Out[26]:
In [27]:
import matplotlib.pyplot as plt
fig, ax = plt.subplots(figsize=(20, 10))
ax.set_ylim(min(0,filteredData1['Value'].min()), 2*filteredData1['Value'].max())
ax.set_title('Indicator Name : ' + indicatorsFilter[0])
ax.plot(filteredData1['Year'], filteredData1['Value'] , 'r--', label=filteredData1['CountryName'].unique())
# Add the legend
legend = plt.legend(loc = 'upper center',
shadow=True,
prop={'weight':'roman','size':'xx-large'})
# Rectangle around the legend
frame = legend.get_frame()
frame.set_facecolor('.95')
plt.show()
In [28]:
import matplotlib.pyplot as plt
fig, ax = plt.subplots(figsize=(20, 10))
# Adjust the lower and upper limit to bring the graph at center
ax.set_ylim(min(0,filteredData2['Value'].min()), 2*filteredData2['Value'].max())
ax.set_title('Indicator Name : ' + indicatorsFilter[0])
ax.plot(filteredData2['Year'], filteredData2['Value'] ,
label=filteredData2['CountryName'].unique(),
color="purple", lw=1, ls='-',
marker='s', markersize=20,
markerfacecolor="yellow", markeredgewidth=4, markeredgecolor="blue")
# Add the legend
legend = plt.legend(loc = 'upper left',
shadow=True,
prop={'weight':'roman','size':'xx-large'})
# Rectangle around the legend
frame = legend.get_frame()
frame.set_facecolor('.95')
plt.show()
In [29]:
from mpl_toolkits.mplot3d import Axes3D
from matplotlib import cm
import matplotlib.pyplot as plt
import numpy as np
countof_angles = 36
countof_radii = 8
# array - radii
array_rad = np.linspace(0.125, 1.0, countof_radii)
# array - angles
array_ang = np.linspace(0, 2*np.pi, countof_angles, endpoint=False)
# repeat all angles per radius
array_ang = np.repeat(array_ang[...,np.newaxis], countof_radii, axis=1)
# from polar (radii, angles) coords to cartesian (x, y) coords
x = np.append(0, (array_rad*np.cos(array_ang)).flatten())
y = np.append(0, (array_rad*np.sin(array_ang)).flatten())
# saddle shaped surface
z = np.sin(-x*y)
fig = plt.figure(figsize=(20,10))
ax = fig.gca(projection='3d')
ax.plot_trisurf(x, y, z, cmap=cm.autumn, linewidth=0.2)
plt.show()
fig.savefig("vis_3d.png")
In [30]:
%matplotlib inline
import numpy as np
import matplotlib.pyplot as plt
n_points = 200
radius = 2 * np.random.rand(n_points)
angles = 2 * (np.pi) * np.random.rand(n_points)
area = 400 * (radius**2) * np.random.rand(n_points)
colors = angles
fig = plt.figure(figsize=(20,10))
ax = plt.subplot(111, polar=True)
c = plt.scatter(angles, radius, c=colors, s=area, cmap=plt.cm.hsv)
c.set_alpha(1.95)
plt.show()
fig.savefig("vis_bubbleplot.png")
In [31]:
np.random.seed(452)
# Three ararys of 100 points each
A1 = np.random.normal(0, 1, 100)
A2 = np.random.normal(0, 2, 100)
A3 = np.random.normal(0, 1.5, 100)
# Concatenate the three arrays
data = [ A1, A2, A3 ]
fig, axes = plt.subplots(nrows=1, ncols=2, figsize=(20, 10))
# Box plot: Notch Shape
bplot1 = axes[1].boxplot(data,
notch=True,
vert=True, # vertical aligmnent
patch_artist=True) # color
# Box plot: Rectangular
bplot2 = axes[0].boxplot(data,
vert=True, # vertical aligmnent
patch_artist=True) # color
colors = ['tomato', 'darkorchid', 'lime']
# more colors here: http://matplotlib.org/examples/color/named_colors.html
for bplot in (bplot1, bplot2):
for patch, color in zip(bplot['boxes'], colors):
patch.set_facecolor(color)
# Grid lines, Xticks, Xlabel, Ylabel
for axis in axes:
axis.yaxis.grid(True)
axis.set_xticks([y for y in range(len(data))], )
axis.set_xlabel('Sample X-Label',fontsize=20)
axis.set_ylabel('Sample Y-Label',fontsize=20)
# Xtick labels
plt.setp(axes, xticks=[y for y in range(len(data))],
xticklabels=['X1', 'X2', 'X3'])
plt.show()
fig.savefig("vis_boxplot.png")
In [ ]: