Scribd
Upload
18 views

Python-Pandas Notes

Pandas can be used to organize data, apply statistical tools, and plot multiple graphs in one figure using different data structures like Series, DataFrames, and Panels. Series are used for 1D data, DataFrames for 2D data like tables defining rows and columns, and Panels for 3D data. Pandas allows reading and writing data from files, plotting graphs, handling missing values, and organizing data.

Uploaded by

Filza Faisal
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
18 views

Python-Pandas Notes

Pandas can be used to organize data, apply statistical tools, and plot multiple graphs in one figure using different data structures like Series, DataFrames, and Panels. Series are used for 1D data, DataFrames for 2D data like tables defining rows and columns, and Panels for 3D data. Pandas allows reading and writing data from files, plotting graphs, handling missing values, and organizing data.

Uploaded by

Filza Faisal
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

#!

/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
PANDAS

Video link: []

Pandas are used for plotting multiple graphs and to interpret


different documents
-how can you organize your data
-applying statistical tools

"""

"""
---------PLOTING MULTIPLE KINDS OF GRAPHS IN ONE
FIGURE-----------------

You can plot mutiple graphs in one go using PANDAS


-three main data structures: Series, data frame, panel
Series: if for 1d data
data frame: numpy(for numbers only) and 2D data
Panel: is for 3D data

"""
import pandas as pd
import matplotlib.pyplot as plt

"""
---------SERIES--------------
"""

x=pd.Series([34774,3463,3547,2346]) #the y-axis you want to draw


#can have list, dict etc
print(x.index) #tells the default index in terms of numbers
#if you want to attribute it to a string, the use the following\

x.index=['Punjab','Sindh','Balochistan','KPK'] #the writings on x-


asis

x.name='Provinces' #to assign the name to your data

print(x.mean) #for mean. also can use median,mode,min,max


print(x.quantile(q=0.5)) #for quantile

fig,axes=plt.subplots(1,3,figsize=(12,5)) #(1,3 are the R,C)

x.plot(ax=axes[0],kind='bar',title='Bar Diagram') #the ax is


defined with axes=0
x.plot(ax=axes[1],kind='line',title='Line Diagram')
x.plot(ax=axes[2],kind='pie',title='Piechart')
######for dictionary########
#you can assign the indexes with the values directly here

z=pd.Series({'a':1,'b':2,'g':3})
print(z) #will give
#a 1
#b 2
#g 3

y=pd.Series([{'a':1,'b':2},[3,4]]) #will print the dictionary at


index 0 and list as 1

#x. plot. bar(x = 'Subject', y = 'Mean', fontsize='9') #another way


of ploting a graph

"""
--------DATA FRAME(MAKING A TABLE(defiing rows and
columns))------------------------------
we can make a table using pandas by creating a dictionary and
defining the variables

Used for a two dimensional list

"""

#first define a data frame and that assigns the colums and the rows
if you first define the column heading and then its contents
df=pd.DataFrame({'Province':['S','P','B','KPK'],'Population':
[1000,2000,3000,4566]})
print(df)
print(df.Population) #to get the content of the population only
print(df.loc[0]) #will tell the information of the first in province
and its corresponding population, in this case: Sindh: 1000
print(df['Province']) #another way of getting a column of elements
df.columns=['Pov','Pops'] #this will change column headings
df.index=['a','b','c','d'] #change index from no. to this

print(df.pop) #to view population column, can be done for pops also
print(df.loc['a']) #city S
#pop 1000 for these
print(df.info()) #for all the info

"""
---------WRITING TO THE FILES--------
"""

#so if I want to write a data to excel I will

df=pd.DataFrame({'Province':['S','P','B','KPK'],'Population':
[1000,2000,3000,4566]})
df_1=df.to_excel('test.xlsx')
df_2=df.to_excel('test.xlsx',index='False') #the index will not
appear

#just to summarize, if you want to read the row use loc[0]


#and if read column, use the heading of that column #df.Population
"""
-------------READING FROM THE FILESS-----------------------
"""
#if you are unable to read your file, then write a file, read it and
then read it again
#or write the whole address p=pd.read_csv('/Users/filzafaisal/
Desktop/pokemon_data.csv')
#find the pathway from excel>file>properties to find the link
address

#if you want to read what you have written


l=pd.read_excel('test.xlsx')
print(l)

#####for general information about data#####


print(l.info())

p=pd.read_csv('pokemon_data.csv') #to open the data file in csv


forma, if you have an xls format, =pd.read_excel('pokemon.xlsx')
#for text use =pd.read_csv('pokemon.txt')
print(p) #to see the entire doc
print(p.loc[0]) #mehtod to see the row you want
print(p.loc['False']) #to see the location of particular value
print(p.head(3)) #to read the top 3 headings
print(p.tail(3)) #to read the bottom 3 headings
print(p.columns) #to print the column headings so you know the names
of the columns
print(p.Name) #to read the elements in a specific column that you
want
print(p[['Name','Attack']]) #if you want to view many columns at
once,just make a list of them

"""
---------MAKE CHANGES IN EXCEL FILE---------
"""

data=pd.read_excel('sample.xlsx')
#then set an index according to the column you want (better if not a
number)

data_new=data.set_index(['Ship
Mode','Customer']).sort_index(level=0) #this made sure that both are
side by side
#to find the values & data within an index
print(data_new.loc['Regular Air'])
#if there were two strings you could have done
print(data_new.loc['Regular Air','Barry French'])

customer_count=data.Customer.value_counts() #apply on original data


#this tells you no. of times customers have been repeated

fig,axes=plt.subplots(1,2,figsize=(12,5))
customer_count.plot(ax=axes[0],kind='bar') #for the customer count
axes[0].set_ylabel('customer') #set y label

"""
-------ADDRESS MISSING NUMBERS-----------
"""

sample2=pd.read_excel('sample2.xlsx')
print(sample2['Unit Price'].isnull()) #if True: means that those
values are null
#so we fill those numbers with 0 by
print(sample2['Unit Price'].fillna(0))

sample3=sample2['Unit Price'].fillna(0) #to have a seperate data


#if you want to fill the data with the previous values
print(sample2['Unit Price'].bfill(0))

#if you want to fill the data with the forward values
print(sample2['Unit Price'].ffill(0))

"""
---------DRAWING GRAPHS OF DATA FRAME(2D FIGS)-----------
"""

#METHOD 1(using series)


#we can use series data to make a series data
y=pd.Series(['khi','lahore'])
z=pd.Series(['sindh','punjab'])
s1=pd.DataFrame([y,z])

s1.index=['a','b']
print(s1)

#METHOD 2(using dictionary)


s2=pd.DataFrame({'Cities':['khi','lahore'],'Age':[30,20]})
#s2=pd.DataFrame({'Cities':pd.Series(['khi','lahore']),'Age':pd.Seri
es([30,20])})
fig,axes=plt.subplots(1,2,figsize=(12,5))
s2.plot(ax=axes[0],x='Cities',y='Age',kind='bar',label='stats')

"""
-----CONVERTING FLOAT TO NUMBERS(by adding new colums)-----
make new columns by editing the existing one
"""
#lambda tells python that we are finding a function
y=lambda x: x**2+1

k=pd.DataFrame({'Cities':pd.Series(['khi','lahore']),'Age':pd.Series
(['30','20'])})
#so here, we will add a new column by lambda
k['NuAge']=k.Age.apply(lambda x: int(x.replace(',','')))

#this will duplicate your existing column and make new one with
edits

"""
--------ORGANIZING DATA------
if you want to organize a particular column in aligned format

"""

k['Cities']=k.Cities.apply(lambda x: x.strip())

#if you want to replace date in index

k.index=pd.date_range('2016-01-01',periods=2) #for the four rows


present in k
print(k) #so here we have dates in the two columns

You might also like