Dimension Reduction with Game of Thrones Battle Dataset

In [383]:

    

# Imports
import numpy as np
import matplotlib as mpl
import matplotlib.cm as cm
import pandas as pd
import sklearn

# Other Imports
from sklearn import preprocessing
from matplotlib import pyplot as plt
from sklearn.preprocessing import StandardScaler
from sklearn.manifold import TSNE

# Inline Plotting
import IPython.html.widgets
%matplotlib inline


    

In [384]:

    

# Load CSV file into Pandas DataFrame
df = pd.read_csv('data/battles.csv')

df.info()


    

    




<class 'pandas.core.frame.DataFrame'>
RangeIndex: 38 entries, 0 to 37
Data columns (total 25 columns):
name                  38 non-null object
year                  38 non-null int64
battle_number         38 non-null int64
attacker_king         36 non-null object
defender_king         35 non-null object
attacker_1            38 non-null object
attacker_2            10 non-null object
attacker_3            3 non-null object
attacker_4            2 non-null object
defender_1            37 non-null object
defender_2            2 non-null object
defender_3            0 non-null float64
defender_4            0 non-null float64
attacker_outcome      37 non-null object
battle_type           37 non-null object
major_death           37 non-null float64
major_capture         37 non-null float64
attacker_size         24 non-null float64
defender_size         19 non-null float64
attacker_commander    37 non-null object
defender_commander    28 non-null object
summer                37 non-null float64
location              37 non-null object
region                38 non-null object
note                  5 non-null object
dtypes: float64(7), int64(2), object(16)
memory usage: 7.5+ KB

In [385]:

    

# Drop Useless Columns
df = df.drop(['name','year','battle_number','attacker_2','attacker_3','attacker_4','defender_2','defender_3', 'defender_4', 'note'], axis=1)

# Drop battles with not enough information
df.dropna(subset=['defender_size', 'attacker_size','attacker_outcome'], axis=0, inplace=True)


    

In [386]:

    

# Reorder columns so outcome is last
df = pd.concat([df.ix[:,:4], df.ix[:,5:], df.attacker_outcome], axis=1)


    

In [387]:

    

# Encode text columns
column_encoders = {}
for c in list(df.columns):
    if df[c].dtype == object:
        le = preprocessing.LabelEncoder()
        df[c] = le.fit_transform(list(df[c]))
        column_encoders[c] = le

df.info()


    

    




<class 'pandas.core.frame.DataFrame'>
Int64Index: 15 entries, 0 to 30
Data columns (total 15 columns):
attacker_king         15 non-null int64
defender_king         15 non-null int64
attacker_1            15 non-null int64
defender_1            15 non-null int64
battle_type           15 non-null int64
major_death           15 non-null float64
major_capture         15 non-null float64
attacker_size         15 non-null float64
defender_size         15 non-null float64
attacker_commander    15 non-null int64
defender_commander    15 non-null int64
summer                15 non-null float64
location              15 non-null int64
region                15 non-null int64
attacker_outcome      15 non-null int64
dtypes: float64(5), int64(10)
memory usage: 1.9 KB

In [388]:

    

# Split Data into X and Y vectors, where Y is Outcome
y = df.ix[:,-1].values
x = df.ix[:,:-1].values

# Standardize features by removing the mean and scaling to unit variance
standard_scaler = StandardScaler()
x_std = standard_scaler.fit_transform(x)


    

In [389]:

    

# t-distributed Stochastic Neighbor Embedding (t-SNE)
# Compress X vector to two dimensions
tsne = TSNE(n_components=2, random_state=0)
x_1d = tsne.fit_transform(x_std)*10000


    

In [390]:

    

# Scatter Plot
plt.figure()

# Create color map and markers for the 2 Outcomes
markers=('s', 'd')
color_map = {0:'red', 1:'blue'}

# Fill Scatter Plot
for idx, cl in enumerate(np.unique(y)):
    plt.scatter(x=x_1d[y==cl,0], y=x_1d[y==cl,1], c=color_map[idx], marker=markers[idx], 
                label=column_encoders['attacker_outcome'].inverse_transform(cl))

# Labels and legend
plt.xlabel('X in t-SNE')
plt.ylabel('Y in t-SNE')
plt.legend(loc='upper left', frameon=True)
plt.title('t-SNE visualization of data')
plt.show()


    

In [ ]: