import pandas as pd
import numpy as np

%matplotlib inline
import matplotlib.pyplot as plt

# setting plot defatult size
%pylab inline
pylab.rcParams['figure.figsize'] = (12, 6)

df = pd.read_csv('data/ind_pop_data.csv')
df.head()

df.values

df.index

#1. by zip lists

list_keys = ['country', 'population']
list_values = [['targaryen', 'stark'], [579, 973]]

zipped = list(zip(list_keys,list_values))
data = dict(zipped)
df_zip = pd.DataFrame(data)
df_zip

df_zip.columns = ['kingdom', 'army'];   # change column labels
df_zip

# 2. by broadcasting

data = {'kingdom':list_values[0], 'babies':list_values[1]}
df_dict = pd.DataFrame(data)
df_dict

# df1 = pd.read_csv(file_messy, header=0, names=new_col_names)
# df2 = pd.read_csv(file_messy, delimiter=' ', header=3, comment='#')

# df2.to_csv(file_clean, index=False)              # Save the DataFrame to a CSV file without the index
# df2.to_excel('file_clean.xlsx', index=False)   # Save the DataFrame to an excel file without the index

df_pop = pd.read_csv('data/ind_pop_data.csv')
df_pop.head()

# example 1

df_c = df_pop[df_pop['CountryCode'] == 'CEB'].loc[:,['TotalPop','UrbanPopRatio']]
df_c.TotalPop = df_c.TotalPop / 1000000
df_c.head()

df_c.plot()

# example 2

df_k = df_pop[df_pop['CountryCode'] == 'KOR'].loc[:,['Year','TotalPop','UrbanPopRatio']]
df_k.TotalPop = df_k.TotalPop / 1000000
df_k.head()

plt.scatter(df_k['Year'], df_k['TotalPop'])
plt.scatter(df_k['Year'], df_k['UrbanPopRatio'])
plt.legend()
plt.title('Increase population in Korea')
plt.xlabel('Year')

# example 3

year_s = pd.Series(list(df_k['Year']))
kor_s = pd.Series(list(df_k['UrbanPopRatio']))
ceb_s = pd.Series(list(df_c['UrbanPopRatio']))

df_new = pd.DataFrame()
df_new['Year'] = year_s
df_new['KOR'] = kor_s
df_new['CEB'] = ceb_s

df_new.head()

df_new.plot(x='Year', y=['KOR','CEB'])

kor_s = pd.Series(list(df_k['TotalPop']))
ceb_s = pd.Series(list(df_c['TotalPop']))

df_new = pd.DataFrame()
df_new['KOR'] = kor_s
df_new['CEB'] = ceb_s

df_new.head()

cols = ['KOR', 'CEB']
df_new[cols].plot(kind='box')

cols = ['KOR', 'CEB']
df_new[cols].plot(kind='box', subplots=True)   # Generate the box plots

df = pd.read_csv('data/worldreport.csv', index_col=0)
df.head()

df.describe()

df['life_exp'].plot(kind='hist', normed=True, bins=30)

# 누적 히스토그램
df['life_exp'].plot(kind='hist', normed=True, cumulative=True, bins=30, range=(35,95))

df['life_exp'].plot(kind='box', figsize=(6,6))

# subplot 분리

fig, axes = plt.subplots(nrows=1, ncols=3)
df_pop.loc[df_pop['Year'] == 1960].plot(ax=axes[0], y='UrbanPopRatio', kind='box', label='UrbanPopRatio 1960')
df_pop.loc[df_pop['Year'] == 1970].plot(ax=axes[1], y='UrbanPopRatio', kind='box', label='UrbanPopRatio 1970')
df_pop.loc[df_pop['Year'] == 1980].plot(ax=axes[2], y='UrbanPopRatio', kind='box', label='UrbanPopRatio 1980')

# df = pd.read_csv(filename, index_col='Date', parse_dates=True)

df_time = pd.read_csv('data/Product_airtime.csv')
df_time.head()

date_list = df_time['PRODUCT_START_TMS']
date_list[0:10]

# Convert date_list into a datetime object
time_format = '%Y-%m-%d %H:%M'
my_datetimes = pd.to_datetime(date_list, format=time_format)
my_datetimes[0:10]

df_time = pd.read_csv('data/Product_airtime.csv', index_col='PRODUCT_START_TMS', parse_dates=True)
df_time.head()

dateRange = pd.date_range('2017/03/01', periods=20, freq='D')
ts = pd.Series(range(len(dateRange)), index=dateRange)
ts[0:10]

ts1 = ts.loc['2017-03-09']
ts1

ts2 = ts.loc['2017-03-05':'2017-03-10']
ts2

sum12 = ts1 + ts2
sum12

ts3 = ts2.reindex(ts.index)
ts3

ts4 = ts2.reindex(ts.index, method='ffill')
ts4

sum23 = ts2 + ts3
sum23

df_time = pd.read_csv('data/Product_airtime.csv', index_col='PRODUCT_START_TMS', parse_dates=True)
df_time.head()

october = df_time['PRODUCT_AIRTIME_MINS']['2012-10']

# Downsample to obtain only the daily highest temperatures in August: august_highs
october_highs = october.resample('D').max()
print(october_highs)

# Rolling means (or moving averages) 

unsmoothed = df_time['PRODUCT_AIRTIME_MINS']['2012-10-01':'2012-10-10']

# Apply a rolling mean with a 24 hour window: smoothed
smoothed = unsmoothed.rolling(window=24).mean()

# Create a new DataFrame with columns smoothed and unsmoothed: august
august = pd.DataFrame({'smoothed':smoothed, 'unsmoothed':unsmoothed})

# Plot both smoothed and unsmoothed data using august.plot().
august.plot()
plt.show()

# Plotting time series, datetime indexing

october_highs.plot()
plt.legend()
plt.xlabel('Date')

october_highs['2012-10-01':'2012-10-10'].plot()
plt.legend()
plt.xlabel('Date')