import pandas as pd
import numpy as np
import seaborn as sns
import matplotlib.pyplot as plt

df_case = pd.read_csv('RAW_global_confirmed_cases.csv')
df_death = pd.read_csv('RAW_global_deaths.csv')

pd.set_option('display.max_columns', None)

df_case.head()

df_case['Country/Region'].value_counts()

def fix_dataframe(df):
    df = df.drop(['Lat', 'Long'], axis=1).groupby('Country/Region').sum()
# 나라 기준으로 묶고 Province당 경우를 합하여 표현한다.

    df = df.transpose()
# 날짜가 인덱스가 되게끔 바꾸어준다.

    df.index.name = 'Date'
    df.reset_index(inplace=True)
    df['Date'] = df['Date'].apply(lambda s : pd.to_datetime(str(s)))

    df.set_index('Date', inplace=True)
    return df

df_case = fix_dataframe(df_case)
df_death = fix_dataframe(df_death)

# 확진자 상위 10개국

ten_cases = df_case.loc[df_case.index[-1]].sort_values(ascending=False)[:10]
ten_cases

sns.barplot(x=ten_cases.index, y=ten_cases)

ten_death = df_death.loc[df_death.index[-1]][ten_cases.index] 
#sort_values 말고 확진자 10개국의 사망자를 확인해보자.
ten_death

# 확진자와 사망자를 동시에 보여주기.
plt.figure(figsize=(7,5))

sns.barplot(x=ten_cases.index, y=ten_cases, color='black')
plt.xticks(rotation=90, size=15)
plt.ylabel('Total Confirmed Cases', size=15)
plt.xlabel('')
plt.title('Total Confirmed Cases (%s)' %ten_cases.name.strftime('%Y-%m-%d'), size=15)

ax = plt.gca()
ax2 = ax.twinx()
ax2.plot(ten_death.index, ten_death, 'r--')
ax2.set_ylabel('Total Deaths', color='red', size=15)
plt.show()

# 미국만 알아보자.
plt.figure(figsize=(10,7))

plt.plot(df_case.index, df_case['US'], 'b-')
plt.ylabel('Confirmed Cases', color='blue')
plt.xlabel('Date')
plt.xlim(right=df_case.index[-1])
plt.ylim(0, df_case['US'].max()*1.1)
plt.title('US' + 'Cases & Deaths')

ax = plt.gca()
ax2 = ax.twinx()
ax2.plot(df_death.index, df_death['US'], 'r--')
ax2.set_ylabel('Deaths', color='red')
ax2.set_ylim(0, df_death['US'].max()*1.3)

plt.show()

# 재미로 일본도

plt.figure(figsize=(6,4))

plt.plot(df_case.index, df_case['Japan'], 'b-')
plt.ylabel('Confirmed Cases', color='blue')
plt.xlabel('Date')
plt.xlim(right=df_case.index[-1])
plt.ylim(0, df_case['Japan'].max()*1.1)
plt.title('US' + 'Cases & Deaths')

ax = plt.gca()
ax2 = ax.twinx()
ax2.plot(df_death.index, df_death['Japan'], 'r--')
ax2.set_ylabel('Deaths', color='red')
ax2.set_ylim(0, df_death['Japan'].max()*1.3)

plt.show()

# 한국
plt.figure(figsize=(6,4))

plt.plot(df_case.index, df_case['Korea, South'], 'b-')
plt.ylabel('Confirmed Cases', color='blue')
plt.xlabel('Date')
plt.xlim(right=df_case.index[-1])
plt.ylim(0, df_case['Korea, South'].max()*1.1)
plt.title('US' + 'Cases & Deaths')

ax = plt.gca()
ax2 = ax.twinx()
ax2.plot(df_death.index, df_death['Korea, South'], 'r--')
ax2.set_ylabel('Deaths', color='red')
ax2.set_ylim(0, df_death['Korea, South'].max()*1.3)

plt.show()

# 누적 말고 일일데이터는?
plt.plot(df_case.index, df_case['Korea, South'].diff(), 'b-')
plt.ylabel('Confirmed Cases', color='blue')
plt.xlabel('Date')
plt.xlim(right=df_case.index[-1])
plt.ylim(bottom = 0)
plt.title('US' + 'Cases & Deaths')

ax = plt.gca()
ax2 = ax.twinx()
ax2.plot(df_death.index, df_death['Korea, South'].diff(), 'r--')
ax2.set_ylabel('Deaths', color='red')
ax2.set_ylim(bottom=0)

plt.show()

df_case.reset_index()[['Date', 'Korea, South']]

df = pd.DataFrame(df_case.reset_index()[['Date', 'Korea, South']].to_numpy(), columns=['ds', 'y'])
df

from math import floor

def train_test_split_df(df, test_size):
    div = floor(df.shape[0] * (1 - test_size))
    return df.loc[:div], df.loc[div + 1:]

train_df, test_df = train_test_split_df(df, 0.1)

train_df.shape

train_df.tail()

test_df.head()

from fbprophet import Prophet

model = Prophet(changepoint_range = 1.0)
model.fit(train_df)

from fbprophet.plot import add_changepoints_to_plot

pred = model.predict(test_df)
model.plot(pred);

model.plot_components(pred);
# 주마다 경향

fig = model.plot(pred)
plt.plot(test_df['ds'], test_df['y'], 'g-', label='actual')
add_changepoints_to_plot(fig.gca(), model, pred)
plt.legend()

from sklearn.metrics import r2_score
print('R2 Score : ', r2_score(test_df['y'], pred['yhat']))

model = Prophet(changepoint_range=1.0)
model.fit(df)
future = model.make_future_dataframe(30)
pred = model.predict(future)
model.plot(pred);

df.loc[26:32]

# 2월 20일 기준으로 급격히 늘어남.

model = Prophet(changepoint_range=1.0)
model.fit(df.loc[:28])
future2 = model.make_future_dataframe(30)
pred = model.predict(future2)
model.plot(pred);

plt.plot(df.loc[:58]['ds'], df.loc[:58]['y'], 'g-')
plt.show()