1)      Read data into a dataframe from the file ‘stock indices.csv’ (which can be downloaded

in BlackBoard> IC208> Lecture Materials> Week 5).

2)      Save data of individual stock index in separate csv files. For example, FTSE Index

would be saved in ‘Uk.csv’ or ‘FTSEALLSHARE.csv’ . DAX30 Index would be saved in ‘Germany.csv’ or ‘DAX30.csv’

3)      Read all stock indices into a data frame from the individual files.

4)      Save data of each stock index in a separate subfolder. For example, FTSE Index would

be saved in subfolder named ‘Uk’ or ‘FTSEALLSHARE’ . DAX30 Index would be saved in subfolder named ‘Germany’ or ‘DAX30’ .

5)      Inspect basic characteristics and summary statistics of the dataframe.

6)      Create a new column, named ‘num’, which has random numbers in interval [0,1)

7)      Drop the ‘num’ column

8)      Rename the column ‘indx’ to ‘price’

9)      Count how many rows in the dataframe is for the UK i.e. country==’Uk’

10)    Create a new dataframe which only includes observations/rows for the UK

11)    Add a column of percentage return of the FTSE All Share index

12)    Add a column of logarithmic return of the FTSE All Share index

13)    Plot FTSE All Share returns over time.

14)    Plot the stock indices of all country in the dataframe in one graph

15)    Create a dataframe that has 4 columns corresponding to logarithm returns of 04 stock indices.

There are many ways to code in Python. The below is only a guide.

1)        Ans:

import os

import pandas as pd

os.chdir(<path>)

data = pd.read_csv('stock indices.csv')

2)        Ans:

import os, glob

os.chdir(<path>)

glob.glob(

for c in countries:

temp_data = data[data['country']==c]

temp_data.to_csv('{}.csv'.format(c), index=False)

3)        Ans:

import os, glob

os.chdir(<path>)

csv_files = glob.glob('*.csv')

data = pd.DataFrame()

for f in csv_files:

temp_data = pd.read_csv(f)

data = data.append(temp_data)

4)        Ans:

data = pd.read_csv('stock indices.csv')

countries = list(data['country'].unique())

for c in countries:

temp_data = data[data['country']==c]

os.mkdir('{}'.format(c))

temp_data.to_csv('{}\{}.csv'.format(c,c), index=False)

5)        Ans:

data.shape

data.size

data.dtypes

data.columns.values

data.sample(10)

data.describe()

6)        Ans:

numbers = np.random.rand(data.shape[0])

data['num'] = numbers

7)        Ans:

data = data.drop(columns=['num'])

8)        Ans:

data = data.rename(columns={'indx':'price'})

9)        Ans:

data[data['country']=='Uk'].shape[0]

#or

mask = (data['country']=='Uk')

data[mask].shape[0]

10)      Ans:

data_uk = data[data['country']=='Uk']

data_uk['datetime'] = pd.to_datetime(data_uk['date'])

data_uk = data_uk.set_index('datetime')

11)      Ans:

data_uk['pct_return'] = data_uk['price'].pct_change()

12)      Ans:

data_uk['log_price'] = np.log(data_uk['price'])

data_uk['log_return'] = data_uk['log_price'].diff(1)

#or

data_uk['log_return']= np.log(data_uk['price']/data_uk['price'].shift(1))

13)      Ans:

data_uk['log_return'].plot()

data_uk[['log_return','pct_return']].plot()

#or

import matplotlib as mpl

from matplotlib import pyplot as plt

fig=plt.figure()

ax1=fig.add_axes([0.1,0.1,0.8,0.8])

ax1.plot(data_uk['log_return'])

ax1.set_xlabel('date')

ax1.set_ylabel('log return')

ax1.set_title('FTSE All Share')

plt.show()

14)      Ans:

new_data = pd.DataFrame()

countries = list(data['country'].unique())

for c in countries:

temp_data = data.loc [data['country']==c,['price','date']] temp_data['datetime'] = pd.to_datetime(temp_data['date']) temp_data = temp_data.set_index('datetime')              new_data['{}'.format(c)] = temp_data['price']

#or

idata = data[['country','date','price']]