Python – Page 2 – Stefan Jaspers

Python

Python Basics – Dictionary

Posted on June 10, 2020June 10, 2020by Stefan

Dictionaries Lists

my_list = ['a','b','c','d']
# index     0   1   2   3

my_list[0]
> 'a'
my_list.index['c']
> 2

my_list = ['a','b','c','d']

# index 0 1 2 3

my_list[0]

> 'a'

my_list.index['c']

> 2

Intro

Keys          Values
----          ------ 
Firm Code     FRD
Process Date  10/14/2019
CUSIP         13246879

Keys Values

---- ------

Firm Code FRD

Process Date 10/14/2019

CUSIP 13246879

Representation of a dictionary = { ‘key-1’ : ‘value-1’ , ‘key-2’ : ‘value-2’ , ‘key-3’ : ‘value-3’ , ‘key-4’ : ‘value-4’ } Creation

# Create an empty dict
my_dict = {}
my_dict
> {}

# native
tickersymbols = {'AAPL':'Apple', 'F':'Ford', 'XON':'Exxon'}
> {'AAPL': 'Apple', 'F': 'Ford', 'XON': 'Exxon'}
# OR by lists
tickersymbols = dict([['AAPL','Apple'],['F','Ford'],['XON','Exxon']])
> {'AAPL': 'Apple', 'F': 'Ford', 'XON': 'Exxon'}

# Create an empty dict

my_dict = {}

my_dict

> {}

# native

tickersymbols = {'AAPL':'Apple', 'F':'Ford', 'XON':'Exxon'}

> {'AAPL': 'Apple', 'F': 'Ford', 'XON': 'Exxon'}

# OR by lists

tickersymbols = dict([['AAPL','Apple'],['F','Ford'],['XON','Exxon']])

> {'AAPL': 'Apple', 'F': 'Ford', 'XON': 'Exxon'}

Handling a dictionary

# Adding
tickersymbols['TSLA'] = 'Tesla'
tickersymbols
> {'AAPL': 'Apple', 'F': 'Ford', 'XON': 'Exxon', 'TSLA': 'Tesla'}

# Update
tickersymbols['TSLA'] = 'Tesla_Inc'
tickersymbols
> {'AAPL': 'Apple', 'F': 'Ford', 'XON': 'Exxon', 'TSLA': 'Tesla_Inc'}

# Deleting an entry
del(tickersymbols['F'])
tickersymbols
> {'AAPL': 'Apple', 'XON': 'Exxon', 'TSLA': 'Tesla_Inc'}

# Adding

tickersymbols['TSLA'] = 'Tesla'

tickersymbols

> {'AAPL': 'Apple', 'F': 'Ford', 'XON': 'Exxon', 'TSLA': 'Tesla'}

# Update

tickersymbols['TSLA'] = 'Tesla_Inc'

tickersymbols

> {'AAPL': 'Apple', 'F': 'Ford', 'XON': 'Exxon', 'TSLA': 'Tesla_Inc'}

# Deleting an entry

del(tickersymbols['F'])

tickersymbols

> {'AAPL': 'Apple', 'XON': 'Exxon', 'TSLA': 'Tesla_Inc'}

Accessing

# Entry in dict
tickersymbols['XON']
> 'Exxon'

# No entry in dict
tickersymbols['IBM']
> KeyError: 'IBM'

# Second method
tickersymbols.get('XON')
> 'Exxon'
tickersymbols.get('IBM')
> (nothing)
tickersymbols.get('IBM','Missing ticker')
> 'Missing ticker'

# Entry in dict

tickersymbols['XON']

> 'Exxon'

# No entry in dict

tickersymbols['IBM']

> KeyError: 'IBM'

# Second method

tickersymbols.get('XON')

> 'Exxon'

tickersymbols.get('IBM')

> (nothing)

tickersymbols.get('IBM','Missing ticker')

> 'Missing ticker'

Examples

from datetime import datetime, timedelta
alphabet_hist = {datetime.datetime(2019, 7, 30, 0, 0): 1228, datetime.datetime(2019, 7, 31, 0, 0): 1218.2, datetime.datetime(2019, 8, 1, 0, 0): 1211.78, datetime.datetime(2019, 8, 2, 0, 0): 1196.32}
closing_price_date = datetime.datetime(2019, 8, 2, 0, 0)

# Closing price for day before
day_before_closing_price_date = closing_price_date - timedelta(days=1)
# Safely print closing price day before or None if it's missing
print(alphabet_hist.get(day_before_closing_price_date))
> 1211.78

# Get day eight weeks in future
future_closing_price_date = closing_price_date + timedelta(weeks=8)
# Safely get value for the future date or the string 'Missing'
print(alphabet_hist.get(future_closing_price_date, 'Missing'))
> Missing

# Print with key
print(alphabet_hist)
> {datetime.datetime(2019, 8, 2, 0, 0): 1196.32, datetime.datetime(2019, 8, 1, 0, 0): 1211.78, datetime.datetime(2019, 7, 31, 0, 0): 1218.2, datetime.datetime(2019, 7, 30, 0, 0): 1228}
# Remove entry
del(alphabet_hist[closing_price_date])
# Print with key deleted
print(alphabet_hist)
> {datetime.datetime(2019, 8, 1, 0, 0): 1211.78, datetime.datetime(2019, 7, 31, 0, 0): 1218.2, datetime.datetime(2019, 7, 30, 0, 0): 1228}

from datetime import datetime, timedelta

alphabet_hist = {datetime.datetime(2019, 7, 30, 0, 0): 1228, datetime.datetime(2019, 7, 31, 0, 0): 1218.2, datetime.datetime(2019, 8, 1, 0, 0): 1211.78, datetime.datetime(2019, 8, 2, 0, 0): 1196.32}

closing_price_date = datetime.datetime(2019, 8, 2, 0, 0)

# Closing price for day before

day_before_closing_price_date = closing_price_date - timedelta(days=1)

# Safely print closing price day before or None if it's missing

print(alphabet_hist.get(day_before_closing_price_date))

> 1211.78

# Get day eight weeks in future

future_closing_price_date = closing_price_date + timedelta(weeks=8)

# Safely get value for the future date or the string 'Missing'

print(alphabet_hist.get(future_closing_price_date, 'Missing'))

> Missing

# Print with key

print(alphabet_hist)

> {datetime.datetime(2019, 8, 2, 0, 0): 1196.32, datetime.datetime(2019, 8, 1, 0, 0): 1211.78, datetime.datetime(2019, 7, 31, 0, 0): 1218.2, datetime.datetime(2019, 7, 30, 0, 0): 1228}

# Remove entry

del(alphabet_hist[closing_price_date])

# Print with key deleted

print(alphabet_hist)

> {datetime.datetime(2019, 8, 1, 0, 0): 1211.78, datetime.datetime(2019, 7, 31, 0, 0): 1218.2, datetime.datetime(2019, 7, 30, 0, 0): 1228}

Python

Python Basics – Datetime

Posted on June 9, 2020June 10, 2020by Stefan

Datetimes – basic Datetime

from datetime import datetime
black_monday = datetim(1987, 10, 19)
> datetime.datetime(1987, 10, 19, 0, 0)

from datetime import datetime

black_monday = datetim(1987, 10, 19)

> datetime.datetime(1987, 10, 19, 0, 0)

Current Datetime

from datetime import datetime
datetime.now()

1 2	from datetime import datetime datetime.now()

Datetime from a string

black_monday_str = "Monday, October 19, 1987, 9:30 am"
format_str = "%A, %B %d, %Y, %I:%M %p"
datetime.datetime.strptime(black_monday_str, format_str)

> datetime.datetime(1987, 10, 19, 9, 30)

black_monday_str = "Monday, October 19, 1987, 9:30 am"

format_str = "%A, %B %d, %Y, %I:%M %p"

datetime.datetime.strptime(black_monday_str, format_str)

> datetime.datetime(1987, 10, 19, 9, 30)

Year: %y ... 01-99
      %Y ... 0000-9999
Month : %b ... Jan-Dec
        %B ... January-December
        %m ... 01-12
Day : %d ... 01-31
Weekday : %a ... Sun-Sat
          %A ... Sunday-Saturday
          %w ... 0-6
Hour : %H ... 00-23
       %I ... 00-12 
Seconds : %S ... 00-59
Microseconds : %f ... 000000-999999
AM/PM : %p

Year: %y ... 01-99

%Y ... 0000-9999

Month : %b ... Jan-Dec

%B ... January-December

%m ... 01-12

Day : %d ... 01-31

Weekday : %a ... Sun-Sat

%A ... Sunday-Saturday

%w ... 0-6

Hour : %H ... 00-23

%I ... 00-12

Seconds : %S ... 00-59

Microseconds : %f ... 000000-999999

AM/PM : %p

"1937-05-10" --> "%Y-%m-%d"
"Friday, 19 May 01" --> "%A, %d %B %y"

1 2	"1937-05-10" --> "%Y-%m-%d" "Friday, 19 May 01" --> "%A, %d %B %y"

String from Datetime

great_depression = datetime.datetime(1929, 10, 29)
great_depression.strftime("%a, %b %d, %Y")
> "Tue, Oct 29, 1929"

great_depression = datetime.datetime(1929, 10, 29)

great_depression.strftime("%a, %b %d, %Y")

> "Tue, Oct 29, 1929"

Datetimes – working with Datetime attributes

# March 10, 2000 Tech Bubble Crash
tech_bubble = datetime(2000, 3, 10)
# year
yr  = tech_bubble.year
# month
mth = tech_bubble.month
# day
day = tech_bubble.day
print(f"Year: {yr}, Month: {mth}, Day: {day}")

> Year: 2000, Month: 3, Day: 10

# March 10, 2000 Tech Bubble Crash

tech_bubble = datetime(2000, 3, 10)

# year

yr = tech_bubble.year

# month

mth = tech_bubble.month

# day

day = tech_bubble.day

print(f"Year: {yr}, Month: {mth}, Day: {day}")

> Year: 2000, Month: 3, Day: 10

Comparing datetime

from datetime import datetime
lehman = datetime.datetime(2008, 9, 15, 0, 0)
morgan_stanley = datetime.datetime(2008, 9, 22, 0, 0)
tarp = datetime.datetime(2008, 10, 3, 0, 0)

# Lehman Brothers before Morgan Stanley?
lehman_first = lehman < morgan_stanley
print(f"It is {lehman_first} that Lehman Brothers declared bankruptcy first.") 
> It is True that Lehman Brothers declared bankruptcy first.

# Goldman Sachs after TARP?
tarp_first = goldman_sachs > tarp
print(f"It is {tarp_first} that TARP was approved first.")
It is False that TARP was approved first.

# Goldman Sachs and Morgan Stanley same day?
same_time = goldman_sachs == morgan_stanley
print(f"It is {same_time} that Morgan Stanley and Goldman Sachs acted simultaneously")
It is True that Morgan Stanley and Goldman Sachs acted simultaneously

from datetime import datetime

lehman = datetime.datetime(2008, 9, 15, 0, 0)

morgan_stanley = datetime.datetime(2008, 9, 22, 0, 0)

tarp = datetime.datetime(2008, 10, 3, 0, 0)

# Lehman Brothers before Morgan Stanley?

lehman_first = lehman < morgan_stanley

print(f"It is {lehman_first} that Lehman Brothers declared bankruptcy first.")

> It is True that Lehman Brothers declared bankruptcy first.

# Goldman Sachs after TARP?

tarp_first = goldman_sachs > tarp

print(f"It is {tarp_first} that TARP was approved first.")

It is False that TARP was approved first.

# Goldman Sachs and Morgan Stanley same day?

same_time = goldman_sachs == morgan_stanley

print(f"It is {same_time} that Morgan Stanley and Goldman Sachs acted simultaneously")

It is True that Morgan Stanley and Goldman Sachs acted simultaneously

Datetime calculation

from datetime import datetime, timedelta
# TARP passed Oct 3 2008
tarp = datetime(2008, 10, 3)

# Seven days before TARP
week_before = tarp - timedelta(days = 7)
# Print week_before
print(week_before)
> 2008-09-26 00:00:00

# One week after TARP
week_after = tarp + timedelta(weeks = 1)
# Print week_after
print(week_after)
> 2008-10-10 00:00:00

# One year after TARP
year_after = tarp + timedelta(weeks = 52)
# Print year_after
print(year_after)
> 2009-10-02 00:00:00

from datetime import datetime, timedelta

# TARP passed Oct 3 2008

tarp = datetime(2008, 10, 3)

# Seven days before TARP

week_before = tarp - timedelta(days = 7)

# Print week_before

print(week_before)

> 2008-09-26 00:00:00

# One week after TARP

week_after = tarp + timedelta(weeks = 1)

# Print week_after

print(week_after)

> 2008-10-10 00:00:00

# One year after TARP

year_after = tarp + timedelta(weeks = 52)

# Print year_after

print(year_after)

> 2009-10-02 00:00:00

Python

Python – Machine Learnining Code Snippets

Posted on June 7, 2020June 7, 2020by Stefan

Feature Engineering Skewed Distribution You’ve been fine-tuning a machine learning model you created for predicting mortgage defaults on a mortgages DataFrame. You’re aiming to eliminate skewness from your numerical variables by using a log transformation. Despite the transformation, one of your columns is still skewed – why do you think so and what is the […]

Python

Python – Data Visualisation Code Snippets

Posted on June 6, 2020June 6, 2020by Stefan

Bar Chart Plot a horizontal bar chart of the data contained in the df Pandas DataFrame.

lab val
0 A 10
1 B 30
2 C 20

lab val

0 A 10

1 B 30

2 C 20

import matplotlib.pyplot as plt
import seaborn as sns
sns.barplot(x="lab", y="val", data=df)
plt.show()

import matplotlib.pyplot as plt

import seaborn as sns

sns.barplot(x="lab", y="val", data=df)

plt.show()

Count Plot Pet owners were surveyed to determine their preference for dogs, cats or both. Visualize the response data contained in the list pets.

import matplotlib.pyplot as plt
import seaborn as sns
pets = ["cats", "cats", "dogs", "both", "dogs", "both", "cats", "cats", "cats", "dogs", "dogs", "dogs", "dogs", "cats", "cats", "both"]
sns.countplot(pets)
plt.show()

import matplotlib.pyplot as plt

import seaborn as sns

pets = ["cats", "cats", "dogs", "both", "dogs", "both", "cats", "cats", "cats", "dogs", "dogs", "dogs", "dogs", "cats", "cats", "both"]

sns.countplot(pets)

plt.show()

Histogram Plot a histogram of the pH variable in the wine_quality […]

Data Interpretation
...
- Python

How to interprete Results – Statistics and ML results

Posted on June 6, 2020June 6, 2020by Stefan

Statistics Statistics – Categorical variable Which measures of center can be used for a variable that records the type of housing that a person has. The variable has three possible values: for free, renting, own. mode is an appropriate measure. Statistics – IQR Consider a set of data: 2, 3, 4, 5, 6, 6, 7, […]

Python

Python Code Snippets – Import and Cleaning

Posted on June 6, 2020June 6, 2020by Stefan

Importing and Cleaning Importing Pickle-files

import pandas as pd
x = pd.read_pickle('data.pkl')

1 2	import pandas as pd x = pd.read_pickle('data.pkl')

CSV Files

# Use only certain columns
import pandas as pd
cols= ['style', 'type', 'price']
wine= pd.read_csv(file_name, usecols= cols)
print(wine.head())

# Use only certain columns

import pandas as pd

cols= ['style', 'type', 'price']

wine= pd.read_csv(file_name, usecols= cols)

print(wine.head())

# - values in the CSV are replaced by NaN
import pandas as pd
candy= pd.read_csv(file_name, na_values= '-')
print(candy.head())

# - values in the CSV are replaced by NaN

import pandas as pd

candy= pd.read_csv(file_name, na_values= '-')

print(candy.head())

TXT Files

data = "data.txt"
file = open(data, mode="r")
text = file.read()
file.close()

data = "data.txt"

file = open(data, mode="r")

text = file.read()

file.close()

# import a txt file with numeric columns to a np_array
import numpy as np
np.loadtxt("data.txt", delimiter="\t", dtype="int")

# import a txt file with numeric columns to a np_array

import numpy as np

np.loadtxt("data.txt", delimiter="\t", dtype="int")

MATLAB file

# transform a matlab file to python dictionnary
import scipy.io
import numpy as np
mat = scipy.io.loadmat('file.mat')
type(mat)

# transform a matlab file to python dictionnary

import scipy.io

import numpy as np

mat = scipy.io.loadmat('file.mat')

type(mat)

Excel File

import pandas as pd
import xlrd
xls = pd.read_excel(url)

import pandas as pd

import xlrd

xls = pd.read_excel(url)

Cleaning

# Split email in two columns
contact.email.str.split('@', expand = True)

1 2	# Split email in two columns contact.email.str.split('@', expand = True)

# Fill all NaN values with 80
df.fillna(80)

# Fill all NaN values with the next value
df.fillna(method = 'ffill')

# Fill all NaN values with 80

df.fillna(80)

# Fill all NaN values with the next value

df.fillna(method = 'ffill')

# Transform a object column to category
user_xp['status'] = user_xp['status'].astype('category')
user_xp.info()

# Transform a object column to category

user_xp['status'] = user_xp['status'].astype('category')

user_xp.info()

Pivoting a dataframe from wide to long format

# rows = store_id , column = type of store
store_wide= store_info.pivot_table(index= 'store_id', columns= 'type', values= 'sales')
print(store_wide)

# rows = store_id , column = type of store

store_wide= store_info.pivot_table(index= 'store_id', columns= 'type', values= 'sales')

print(store_wide)

# given a Name Height Weight table --> transform it to Name Variable value table
import pandas as pd
pd.melt(measure, id_vars = 'Name', value_vars = 'Height')

# given a Name Height Weight table --> transform it to Name Variable value table

import pandas as pd

pd.melt(measure, id_vars = 'Name', value_vars = 'Height')

# given a table with store-2018-2019-2020
# transform it to store- year - sales
store_long= store_sales.melt(id_vars= 'store', var_name= 'year', value_name= 'sales')
print(store_long.head())

# given a table with store-2018-2019-2020

# transform it to store- year - sales

store_long= store_sales.melt(id_vars= 'store', var_name= 'year', value_name= 'sales')

print(store_long.head())

Python

Python Code – Basic Programming

Posted on June 4, 2020June 6, 2020by Stefan

Lists

packages = ['numpy','pandas','scipy']
for i in packages:
  print(i)

packages = ['numpy','pandas','scipy']

for i in packages:

print(i)

import pandas as pd
list_keys = ['a','b']
list_values = [['x','y'],[1,2]]
all = list(zip(list_keys, list_values))

> all
> [('a',['x','y']),('b',[1,2])]

import pandas as pd

list_keys = ['a','b']

list_values = [['x','y'],[1,2]]

all = list(zip(list_keys, list_values))

> all

> [('a',['x','y']),('b',[1,2])]

[i for i in range(5) if i > 2]
> [3, 4]
[s.upper() for s in ['hallo','aarde']]
> ['HALLO','AARDE']

[i for i in range(5) if i > 2]

> [3, 4]

[s.upper() for s in ['hallo','aarde']]

> ['HALLO','AARDE']

Dictionaries

d = { 'appel': 1, 'peer': 2, 'kokos': 3 }
d['banaan'] = 4

> d
> { 'appel': 1, 'peer': 2, 'kokos': 3, 'banaan': 4 }

d = { 'appel': 1, 'peer': 2, 'kokos': 3 }

d['banaan'] = 4

> d

> { 'appel': 1, 'peer': 2, 'kokos': 3, 'banaan': 4 }

d= {'een': 1, 'twee': 2, 'drie': 3, 'vier': 4} 
> d['twee']
> 2

d= {'een': 1, 'twee': 2, 'drie': 3, 'vier': 4}

> d['twee']

> 2

Arrays numpy

import numpy as np
arr = np.array([1,2,3])

> type(arr)
> numpy.ndarray

import numpy as np

arr = np.array([1,2,3])

> type(arr)

> numpy.ndarray

import numpy as np
x = np.array([1, 2, 3, 4])
y = 3

> x == y
> array([False, False, True, False])

import numpy as np

x = np.array([1, 2, 3, 4])

y = 3

> x == y

> array([False, False, True, False])

import numpy as np
arr = np.array([1, 2, 3, 4, 5])

> arr[arr >3] 
> [4, 5]

import numpy as np

arr = np.array([1, 2, 3, 4, 5])

> arr[arr >3]

> [4, 5]

Iterators

w = 'python'
w_iterator = iter(w)

> next(w_iterator)
> 'p'

w = 'python'

w_iterator = iter(w)

> next(w_iterator)

> 'p'

Lambda Function

f = lambda x, y: x * y

> f(3,3)
> 9

f = lambda x, y: x * y

> f(3,3)

> 9

Functions

def add(a, b= 1):
  return a + b

> add(14)
> 15

def add(a, b= 1):

return a + b

> add(14)

> 15

def double_number(n):
  return n * 2 
double_number(4)

def double_number(n):

return n * 2

double_number(4)

def subtract(n, p=1):
  return n - p

> subtract(4) 
> 3

def subtract(n, p=1):

return n - p

> subtract(4)

> 3

def to_int(x):
  try:
    return int(x)
  except ValueError:
    print("No conversion possible")

> to_int("")
> No conversion possible

def to_int(x):

try:

return int(x)

except ValueError:

print("No conversion possible")

> to_int("")

> No conversion possible

def f():
   """returns two"""
   return 2

> help(f)

Help on function f in module __main__:
f()
   returns two

def f():

"""returns two"""

return 2

> help(f)

Help on function f in module __main__:

f()

returns two

from pytest import *
import numpy as np

def test_addition():
   assert 1 + 1 == 2

> show_test_output(test_addition)

====== test session starts ========
platform linux -- Python 3.6, ...
rootdir: /tmp/tmpxxxx
collected 1 item
test_file.py .

====== 1 passed in 0.01 seconds ==

from pytest import *

import numpy as np

def test_addition():

assert 1 + 1 == 2

> show_test_output(test_addition)

====== test session starts ========

platform linux -- Python 3.6, ...

rootdir: /tmp/tmpxxxx

collected 1 item

test_file.py .

====== 1 passed in 0.01 seconds ==

Object Oriented Programming

class Snoep:
  smaak = 'zoet'
  def __init__(self, name):
    self.name = name

> c = Snoep('Chocolade')
> c.name
> 'Chocolade'

class Snoep:

smaak = 'zoet'

def __init__(self, name):

self.name = name

> c = Snoep('Chocolade')

> c.name

> 'Chocolade'

class hond:
  def woef(self):
    return 'woef'

> milou = hond()
> milou.woef()
> 'woef'

class hond:

def woef(self):

return 'woef'

> milou = hond()

> milou.woef()

> 'woef'