Tag: Python Tutorial

Matplotlib is a Python library used to create charts and graphs.

Installation

$ sudo python3.6 -m pip install pandas

Simple Line plot

import codecademylib  
from matplotlib import pyplot as plt

days = [0, 1, 2, 3, 4,5,6]  
money_spent = [10, 12, 12, 10, 14,22,24]  
plt.plot(days, money_spent)  
plt.show()  

multiple line plots displayed on the same set of axes.

import codecademylib  
from matplotlib import pyplot as plt

time = [0, 1, 2, 3, 4]  
revenue = [200, 400, 650, 800, 850]  
costs = [150, 500, 550, 550, 560]

plt.plot(time, revenue)  
plt.plot(time, costs)  
plt.show()  

Plot options

import codecademylib  
from matplotlib import pyplot as plt

time = [0, 1, 2, 3, 4]  
revenue = [200, 400, 650, 800, 850]  
costs = [150, 500, 550, 550, 560]

plt.plot(time,revenue,color='purple',linestyle='--')  
# marker='o' is circle; s is squar , * is star
plt.plot(time,costs,color='#82edc9',marker='s')  
plt.show()  

special

zoom in and zoom out

# first wto are x coordinates and last two and y coordinates
plt.axis([0, 12, 2900, 3100])  
#Labelling the axis
plt.xlabel('Time of day')  
plt.ylabel('Happiness Rating (out of 10)')  
plt.title('My Self-Reported Happiness While Awake')  

Creating subplots

he command plt.subplot(2, 3, 4) would create “Subplot 4” from the figure above.

Any plt.plot that comes after plt.subplot will create a line plot in the specified subplot

import codecademylib  
from matplotlib import pyplot as plt

x = range(7)  
straight_line = [0, 1, 2, 3, 4, 5, 6]  
parabola = [0, 1, 4, 9, 16, 25, 36]  
cubic = [0, 1, 8, 27, 64, 125, 216]

# Subplot 1
plt.subplot(2, 1, 1)  
plt.plot(x, straight_line)

# Subplot 2
plt.subplot(2, 2, 3)  
plt.plot(x, parabola)

# Subplot 3
plt.subplot(2, 2, 4)  
plt.plot(x, cubic)

plt.subplots_adjust(wspace=0.35, bottom=0.2)

plt.show()  

Adding legends

plt.legend(['Hyrule', 'Kakariko','Gerudo Valley'],loc=8)  

specific ticks

ax.set_xticks([1, 2, 4])  
ax.set_yticks([0.1, 0.6, 0.8])  
ax.set_yticklabels(['10%', '60%', '80%'])  

Working example

from matplotlib import pyplot as plt

month_names = ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep","Oct", "Nov", "Dec"]

months = range(12)  
conversion = [0.05, 0.08, 0.18, 0.28, 0.4, 0.66, 0.74, 0.78, 0.8, 0.81, 0.85, 0.85]

plt.xlabel("Months")  
plt.ylabel("Conversion")

plt.plot(months, conversion)

# Your work here
ax = plt.subplot()  
ax.set_xticks(months)  
ax.set_xticklabels(month_names)

ax.set_yticks([0.10, 0.25, 0.5, 0.75])  
ax.set_yticklabels(['10%', '25%', '50%','75%'])

plt.show()  

Axis lables

plt.xlabel('Time of day')  
plt.ylabel('Happiness Rating (out of 10)')  
plt.title('My Self-Reported Happiness While Awake')  
plt.show()  

Showing legends

plt.legend(['parabola', 'cubic'], loc=6)  
plt.show()  

saving fig

plt.figure(figsize=(7, 3))  
plt.plot(years, power_generated)  
plt.savefig('power_generated.png')  

Different type of graphs

bargraph

from matplotlib import pyplot as plt

drinks = ["cappuccino", "latte", "chai", "americano", "mocha", "espresso"]  
sales =  [91, 76, 56, 66, 52, 27]

plt.bar(range(len(drinks)), sales)

#create your ax object here
ax = plt.subplot()  
ax.set_xticks([0, 1, 2, 3, 4, 5])  
ax.set_xticklabels(["cappuccino", "latte", "chai", "americano", "mocha", "espresso"],  
rotation=90)  
plt.show()  

adjecent bargraph

from matplotlib import pyplot as plt

drinks = ["cappuccino", "latte", "chai", "americano", "mocha", "espresso"]  
sales1 = [91, 76, 56, 66, 52, 27]  
sales2 = [65, 82, 36, 68, 38, 40]

#Paste the x_values code here
n = 1  # This is our first dataset (out of 2)  
t = 2 # Number of dataset  
d = 6 # Number of sets of bars  
w = 0.8 # Width of each bar  
store1_x = [t*element + w*n for element  
             in range(d)]

plt.bar(store1_x, sales1)  
#Paste the x_values code here
n = 2  # This is our second dataset (out of 2)  
t = 2 # Number of dataset  
d = 6 # Number of sets of bars  
w = 0.8 # Width of each bar  
store2_x = [t*element + w*n for element  
             in range(d)]

plt.bar(store2_x, sales2)

plt.show()  

Creating range

from matplotlib import pyplot as plt

months = range(12)  
month_names = ["Jan", "Feb", "Mar", "Apr", "May", "Jun", "Jul", "Aug", "Sep", "Oct", "Nov", "Dec"]  
revenue = [16000, 14000, 17500, 19500, 21500, 21500, 22000, 23000, 20000, 19500, 18000, 16500]

#your work here
ax = plt.subplot()  
ax.set_xticks(months)  
ax.set_xticklabels(month_names)

y_upper = [1.1*i for i in revenue]  
y_lower = [0.9*i for i in revenue]  
plt.fill_between(range(len(months)), y_lower, y_upper, alpha=0.2) 

plt.plot(range(len(months)),revenue)  
plt.show()  

Pie charts

from matplotlib import pyplot as plt

payment_method_names = ["Card Swipe", "Cash", "Apple Pay", "Other"]  
payment_method_freqs = [270, 77, 32, 11]

plt.pie(payment_method_freqs, autopct="%0.1f%%")  
plt.axis('equal')  
plt.legend(payment_method_names)

plt.show()  

Histogram

a = normal(loc=64, scale=2, size=10000)  
b = normal(loc=70, scale=2, size=100000)

plt.hist(a, range=(55, 75), bins=20, alpha=0.5, normed=True)  
plt.hist(b, range=(55, 75), bins=20, alpha=0.5, normed=True)  
plt.show()  

Please refer to matplotlib tutorial for further information.

Pandas is a Python module for working with tabular data (i.e., data in a table with rows and columns). Tabular data has a lot of the same functionality as SQL or Excel, but Pandas adds the power of Python.

• A DataFrame is an object that stores data as rows and columns. You can think of a DataFrame as a spreadsheet or as a SQL table
• You can manually create a DataFrame or fill it with data from a CSV, an Excel spreadsheet, or a SQL query.
• DataFrames have rows and columns. Each column has a name, which is a string.
• Each row has an index, which is an integer.
• DataFrames can contain many different data types: strings, ints, floats, tuples, etc.

Please refer to how to install pandas to install it for python3.6

Pandas have following three data structures

1. Series
2. DataFrames
3. Panels

we mostly will work with DataFrames. series can be considered as subset of DataFrames.

Creating Series and DataFrames

my_list=[1,3,5,6,8,11,18]
s1 = pd.Series(my_list)
s2 = pd.Series([1,3,5,6,8,11,18])
s3 = pd.Series(np.arange(100))

s4=pd.DataFrame(my_list)
df2 = pd.DataFrame(np.random.randint(low=0, high=10, size=(5, 4)),columns=['a', 'b', 'c', 'd'])

>>> type(my_list)
<class 'list'>
>>> type(s1)
<class 'pandas.core.series.Series'>
>>> type(s4)
<class 'pandas.core.frame.DataFrame'>
# Creating DataFrame using dict
sales = [{'account': 'Jones LLC', 'Jun': 150, 'Jul': 200, 'Aug': 140},
{'account': 'Alpha Co', 'Jun': 200, 'Jul': 210, 'Aug': 215},
{'account': 'Blue Inc', 'Jun': 50, 'Jul': 90, 'Aug': 95 }]
df = pd.DataFrame(sales)
#
sales = {'account': ['Jones LLC', 'Alpha Co', 'Blue Inc'],
'Jun': [111, 222, 55],
'Jul': [222, 210, 90],
'Aug': [140, 215, 95]}
df = pd.DataFrame.from_dict(sales)

>>> df.dtypes
Feb int64
Jan int64
Mar int64
account object
dtype: object

df2 = pd.DataFrame(data,columns=['Store ID','Location','Number of Employees'])  
print df2  

Reading from and writing into CSV files

df = pd.read_csv('my-csv-file.csv')
df.to_csv('new-csv-file.csv')
#displaying top 5 rows
df.head()
#displaying top 10 rows
df.head(10)
#statistical information about df
df.info()

df columns can be accessed by two ways

df['age']  
df.age  

type of this data


print(type(df['age'])
print(type(df.age)
<class 'pandas.core.series.Series'>

Accessing rows

print(df.loc[2]) #Python is zero indexed  
df.loc[3:7]  
df.loc[:4]  
df.loc[2:]  

Creating sub dataframe

# create using specific columns
df2 =df[['clinic_north','clinic_south']]  
# create using certian conditon
df2=df[df.month == 'january']  
df2=df[df.age > 27]  
df2=df[df.city != 'Mumbai']  
# using multyiple conditions / In Python, | means "or" and & means "and".
df2 = df[(df.age < 30) | (df.name == 'Martha Jones')] # (curly brackets are must)  
df2 = df[df.month.isin(['January','February','March'])]  
# Reset index to fetch correct row number
df2.reset_index(drop=True)  

Modifying DataFrames

#adding a column
df['height']=[1,2,3,4,5,5]  
#adding same value column
df['In Stock?'] = True  
df['In Stock?'] = 'Yes'  
#New column derived from other columns
df['Revenue'] = df.Price - df['Cost to Manufacture']  
#changing cap
from strings import lower  
df['Lowercase Name'] = df['Name'].apply(lower)  

Using lambda function to modify

get_last_name = lambda x: x.split(' ')[-1]  
df['last_name'] = df.name.apply(get_last_name)  

Columns can be renamed using

df.columns = ['ID', 'Title','Category','Year Released','Rating']  
#selective renaming
df.rename(columns={'name': 'First Name','age': 'Age'},inplace=True)  

Using rename with only the columns keyword will create a new DataFrame, leaving your original DataFrame unchanged. That’s why we also passed in the keyword argument inplace=True. Using inplace=True lets us edit the original DataFrame.

wokring with rows amd columns

import pandas as pd

orders = pd.read_csv('shoefly.csv')

print orders.head(5)

orders['shoe_source'] = orders.shoe_material.apply(lambda x:   
                            'animal' if x == 'leather'else 'vegan')

orders['salutation'] = orders.apply(lambda row:   
                                    'Dear Mr. ' + row['last_name']
                                    if row['gender'] == 'male'
                                    else 'Dear Ms. ' + row['last_name'],
                                    axis=1)

Aggregations

• mean Average of all values in column
• std Standard deviation
• median Median
• max Maximum value in column
• min Minimum value in column
• count Number of values in column
• nunique Number of unique values in column
• unique List of unique values in column

num_colors =orders.price.max()  
num_colors =orders.shoe_color.nunique()  
#groupby
df.groupby('column1').column2.measurement()  
#example
grades = df.groupby('student').grade.mean()  
pricey_shoes =orders.groupby('shoe_type').price.max()  
#groupby with muiple columns
shoe_counts = orders.groupby(['shoe_type','shoe_color']).id.count().reset_index()  

Sometimes, the operation that you want to perform is more complicated than mean or count. In those cases, you can use the apply method and lambda functions, just like we did for individual column operations. Note that the input to our lambda function will always be a list of values.

cheap_shoes = orders.groupby('shoe_color').price.apply(lambda x: np.percentile(x,25)).reset_index()  
print(cheap_shoes)

Pivots using dataframe

import pandas as pd

user_visits = pd.read_csv('page_visits.csv')  
print(user_visits.head())

click_source=user_visits.groupby('utm_source').id.count().reset_index()  
print(click_source)

click_source_by_month=user_visits.groupby(['utm_source','month']).id.count().reset_index()  
print(click_source_by_month)

click_source_by_month_pivot = click_source_by_month.pivot(index='utm_source',columns='month',values='id').reset_index()

print(click_source_by_month_pivot)  

Merge two dataframes

#this will happen if both dataframes have one column common
sales_vs_targets = pd.merge(sales,targets)  
#another way to merge
new_df = orders.merge(customers)  
#multile merge
new_df = orders.merge(customers).merge(products)  
#merge of columns name do not match
orders_products = pd.merge(orders,products.rename(columns={'id': 'product_id'}))  
#another way to merge if columns name do not match
orders_products = pd.merge(orders,products, left_on='product_id', right_on='id', suffixes=['_orders','_products'])  
#outermerge
pd.merge(company_a, company_b, how='outer')  
#left
store_a_b_left = pd.merge(store_a, store_b, how='left')  
#right
store_a_b_right = pd.merge(store_a, store_b, how='right')  
#concatenation
pd.concat([df1, df2])  

Query DataFrame

#simple condition
crushing_it = sales_vs_targets[sales_vs_targets.revenue > sales_vs_targets.target]  
#multiple conditions
results= all_data[(all_data.revenue > all_data.target) &(all_data.women>all_data.men)]  
print(results)  

NumPy, which stands for Numerical Python.

NumPy has many uses including:

• Efficiently working with many numbers at once
• Generating random numbers
• Performing many different numerical functions (i.e., calculating sin, cos, tan, mean, median, etc.)

Importing numpy

This is an universal way of importing NumPy and using np

import numpy as np  

NumPy Arrays

A NumPy array is a special type of list.
Each item can be of any type (strings, numbers, or even other arrays). You can even have different types of items in the same array.
Its best suited for numbers as it gives extra power for mathematical operations

import numpy as np

test_1 = np.array([92, 94, 88, 91, 87])  
print(test_1)  
print(type(test_1))

my_list = [1, 2, 3, 4, 5, 6]  
my_array = np.array(my_list)

print(my_list)  
print(type(my_list))  
print(my_array)  
print(type(my_array))  

Output is as bellow

[92 94 88 91 87]
<type 'numpy.ndarray'>  
[1, 2, 3, 4, 5, 6]
<type 'list'>  
[1 2 3 4 5 6]
<type 'numpy.ndarray'>  

reading from file

csv_array = np.genfromtxt('sample.csv', delimiter=',')  

NumPy arrays are more efficient than lists. One reason is that they allow you to do element-wise operations.

# With a list
l = [1, 2, 3, 4, 5]  
l_plus_3 = []  
for i in range(len(l)):  
    l_plus_3.append(l[i] + 3)
# With an array
a = np.array(l)  
a_plus_3 = a + 3  
>>> np.sqrt(a)
array([ 1, 1.41421356, 1.73205081, 2, 2.23606798, 2.44948974])  

Statistics with NumPy

Lets is consider following data set

import numpy as np

water_height = np.array([4.01, 4.03, 4.27, 4.29, 4.19,  
                         4.15, 4.16, 4.23, 4.29, 4.19,
                         4.00, 4.22, 4.25, 4.19, 4.10,
                         4.14, 4.03, 4.23, 4.08, 14.20,
                         14.03, 11.20, 8.19, 6.18, 4.04,
                         4.08, 4.11, 4.23, 3.99, 4.23])
# Calculate mean
np.mean(water_height)  
# Sort np array
np.sort(water_height)  
# Find median
np.median(water_height)  
# Find percentile value
np.percentile(water_height, 75)  
# Find standard Deviation
np.std(water_height)  
# Percentage of values greater than or equal to 4
np.mean(water_height >= 4)  

This works well for single dimensional array. Lets look at it how it looks for two dimensional array

ring_toss = np.array([[1, 0, 0],  
                          [0, 0, 1], 
                          [1, 0, 1]])
np.mean(ring_toss)  
0.44444444444444442  
# To find the means of each interior array, we specify axis 1 (the "rows"):
np.mean(ring_toss, axis=1)  
# To find the means of each index position, we specify axis 0 (the "columns"):
np.mean(ring_toss, axis=0)