%matplotlib inline
import numpy as np
import matplotlib.pyplot as plt
plt.style.use('seaborn-whitegrid')
import pandas as pd
import ipywidgets as widgets

np.random.seed(1999)
 
N = 100
mydata = {}
mydata['id'] = range(N)
income = np.exp(np.random.normal(size=N))
consumption = np.sqrt(income)

dt = pd.DataFrame(mydata)
dt.head()

# write your code here
dt.head()

# write your code here

# write your code here
dt.head()

# write your code here

# write your code here
# dt['consumption'].mean() # <- compare with answer

# write your code here
# dt['consumption'].mean() # <- compare with answer

print(f'before: {dt.shape[0]} observations, {dt.shape[1]} variables')
# write your code here
print(f'after: {dt.shape[0]} observations, {dt.shape[1]} variables')

# write your code
dt.head()

def assets_row_by_row(x,R,Y):
    return 0 # wrong line
    
def assets_all_at_once(income,consumption,R,Y):
    return 0 # wrong line

def assets_adj(assets,R,Y):
    # missing lines
    pass

R = 1.2 # return rate
Y = 1 # income
try:
    dt['assets_1'] = R*(dt['inc']-dt['con'])+Y
    dt['assets_2'] = dt.apply(assets_row_by_row,axis=1,args=(R,Y))
    dt['assets_3'] = assets_all_at_once(dt['inc'].values(),dt['con'].values(),R,Y)
    dt['assets_4'] = dt['inc']-dt['con']
    assets_adj(dt['assets_4'],R,Y)
except:
    print('failed')
dt.head()    

def _plot_timeseries(dataframe, variable, years):
    
    fig = plt.figure(dpi=100)
    ax = fig.add_subplot(1,1,1)
    
    dataframe.loc[:,['year']] = pd.to_numeric(dataframe['year'])
    I = (dataframe['year'] >= years[0]) & (dataframe['year'] <= years[1])
    
    x = dataframe.loc[I,'year']
    y = dataframe.loc[I,variable]
    ax.plot(x,y)
    
    ax.set_xticks(list(range(years[0], years[1] + 1, 5)))    
    
def plot_timeseries(dataframe):
    
    widgets.interact(_plot_timeseries, 
    dataframe = widgets.fixed(dataframe),
    variable = widgets.Dropdown(
        description='variable', 
        options=['Y','C','G','I','X','M'], 
        value='Y'),
    years=widgets.IntRangeSlider(
        description="years",
        min=1966,
        max=2018,
        value=[1980, 2018],
        continuous_update=False,
    )                 
); 

rename_dict = {}
rename_dict['P.1 Output'] = 'Y'
rename_dict['P.3 Final consumption expenditure'] = 'C'
rename_dict['P.3 Government consumption expenditure'] = 'G'
rename_dict['P.5 Gross capital formation'] = 'I'
rename_dict['P.6 Export of goods and services'] = 'X'
rename_dict['P.7 Import of goods and services'] = 'M'

# a. load data set
# nah1 = ?

# b. rename variables

# c. remove rows where Y is nan

# d. correct year column data
# hint, nah1.loc[J,['year']] = nah1.loc[I,['year']].values

# e. only keep rows with '2010-prices, chained values'

# f. only keep renamed variables

# g. interactive plot
# plot_timeseries(nan)

plot_timeseries(nah1)