First EDA Project

import random 

random.seed(1)
pokemon = pd.read_csv("./input/pokemon.csv")
combat = pd.read_csv("./input/combats.csv")

pokemon.head() #

combat.head()

print("pokemon: " + str(pokemon.shape))
print("combat: " + str(combat.shape))

combat.isnull().sum()

pokemon.isnull().sum()

print("==== Name = NaN, 누구냐 넌 ==== ")

if pokemon[pokemon['Name'].isnull()].size != 0:
    who = pokemon[pokemon['Name'].isnull()]
    print(who, '\n')

    print('앞 포켓몬 : ', pokemon['Name'][int(who.index[0])-1])
    print('뒤 포켓몬 : ', pokemon['Name'][int(who.index[0])+1])
    pokemon[pokemon['#'] ==  63]
else:
    print("이름이 비어있는 포켓몬이 없습니다.")
    # who.Name = "Primeape"

pokemon.loc[(pokemon['#'] ==  63), 'Name'] = "Primeape"
pokemon[pokemon['#'] ==  63]

totalWins = combat.Winner.value_counts() # 전체 승리
countWins = combat.groupby('Winner').count()

countFirst = combat.groupby('First_pokemon').count()
countSecond = combat.groupby('Second_pokemon').count()

print("Count by first winner shape: " + str(countFirst.shape))
print("Count by second winner shape: " + str(countSecond.shape))
print("Total Wins shape : " + str(totalWins.shape))

find_pokemon= np.setdiff1d(countFirst.index.values, countWins.index.values)-1
found = pokemon.iloc[find_pokemon[0],]
print(found)

mergedResult[np.isfinite(mergedResult['Win Percentage'])].sort_values(by = ['Win Percentage']).head(10)

mergedResult[np.isfinite(mergedResult['Win Percentage'])].sort_values(by = ['Win Percentage'], ascending = False ).head(10)

import matplotlib.pyplot as plt
ax = sns.countplot(x="Type 1", data=mergedResult)
plt.xticks(rotation= 90)
plt.xlabel('Type 1')
plt.ylabel('Total ')
plt.title("Pokemon MainType")

ax = sns.countplot(x="Type 2", data=mergedResult)
plt.xticks(rotation= 90)
plt.xlabel('Type 2')
plt.ylabel('Total ')
plt.title("Pokemon SubType")

mergedResult.groupby('Type 1').agg({"Win Percentage": "mean"}).sort_values(by = "Win Percentage")

col = ['Type 1','HP', 'Attack', 'Defense', 'Sp. Atk', 'Sp. Def', 'Speed', 'Win Percentage']
sns.pairplot(mergedResult.loc[:,col].dropna())

# 레퍼런스 https://datascience.stackexchange.com/questions/10459/calculation-and-visualization-of-correlation-matrix-with-pandas
def correlation_matrix(df):
    from matplotlib import cm as cm

    fig = plt.figure()
    ax1 = fig.add_subplot(111)
    cmap = cm.get_cmap('jet', 50)
    cax = ax1.imshow(df.corr(), interpolation="nearest", cmap=cmap)
    ax1.grid(True)
    plt.title('relations')
    labels=['Type 1','HP', 'Attack', 'Defense', 'Sp. Atk', 'Sp. Def', 'Speed', 'Win %']
    ax1.set_xticklabels(labels,fontsize=7)
    ax1.set_yticklabels(labels,fontsize=7)
    fig.colorbar(cax, ticks=[0.00,.05,.10,.15,.20,.25,.30,.35,.40,.45,.50,.55,.60,.65,.70,.75,.8,.85,.90,.95,1])
    plt.show()

correlation_matrix(mergedResult.loc[:,col])

sns.regplot(x="Speed", y="Win Percentage", data=mergedResult, logistic=True).set_title("Speed : Win Percentage")
sns.lmplot(x="Speed", y="Win Percentage", data=mergedResult, hue = 'Type 1',  logistic=True)