Module 2 - Tutorial - Explore Data

# import libraries

import pandas as pd

import numpy as np

print(pd.__version__)

# load csv file data with headers

#location = "smallgradesh.csv" # uncomment this line for Jupyter Notebooks

location = smallgradesh.csv # comment out this line for Jyputer Notebook

df = pd.read_csv(location)

df.head()

# df.head?

df.tail(3)

# load data without headers

# location2 = "smallgrades.csv" # uncomment this in Jupyter Notebook

location2 = smallgrades.csv # comment out this in Jupyter Notebook

df_nohead = pd.read_csv(location2, header=None) # try w/o header=None

df_nohead.head()

# add headers during data load

df_during = pd.read_csv(location2, names=['Name', 'Grade'])

df_during.head()

# add headers after data load

df_nohead.columns = ['Name', 'Grade']

df_nohead.head()

# create data from scratch

names = ['Bob','Jessica','Mary','John','Mel']

grades = [76,95,77,78,99]

GradeList = list(zip(names,grades))

GradeList

# export csv files

df = pd.DataFrame(data = GradeList, columns=['Names','Grades'])

df.to_csv('studentgrades.csv',index=False,header=False)

df

# import Excel file

location = gradedata.xlsx

# location = "gradedata.xlsx"

df = pd.read_excel(location, engine='openpyxl') # overwrites the info from the df variable in the examples above

df.head()

# save dataframe as Excel file

# using GradeList from above

names = ['Bob','Jessica','Mary','John','Mel']

grades = [76,95,77,78,99]

GradeList = list(zip(names,grades))

df = pd.DataFrame(data = GradeList, columns=['Names','Grades'])

writer = pd.ExcelWriter('dataframe_01.xlsx')

df.to_excel(writer, sheet_name='Sheet1')

writer.save()

# multiple sheets

df = pd.DataFrame(data = GradeList, columns=['Names','Grades'])

writer = pd.ExcelWriter('dataframe_02.xlsx')

df.to_excel(writer, sheet_name='Sheet1')

df.to_excel(writer, sheet_name='Sheet2')

writer.save()

# load gradedata.csv file

# location = "gradedata.csv"

location = gradedata.csv

df = pd.read_csv(location)

df.head()

# show the number of rows and columns in a dataframe

df.shape

# show the column names in the dataset

df.columns

# or print(df.columns)

# show the data type of each column

df.dtypes

# number of non-NA values

# len(df) would count rows including NA values

df.count()

df['hours'].max()

df['hours'].min()

df['hours'].sum()

df['hours'].mean()

df['hours'].median()

df['hours'].mode()

df['hours'].value_counts()

# standard deviation

df['hours'].std()

# descriptive statistics

df.describe()

# doesn't do anything on its own

df.groupby('gender')

# add a stats or math function to it

df['hours'].groupby(df['gender']).mean()

# mean of multiple columns

# df[['hours', 'exercise']].groupby(df['gender']).mean()

# use two columns to groupby

df.groupby(['gender', 'age']).mean()

# pivot table default function is mean

pd.pivot_table(df, index=['gender'])

pd.pivot_table(df, values=['hours'], index=['gender'])

# unique values in a column

df['age'].unique()

# find missing values

filename = gradedatamissing.csv # comment out this in Jupyter Notebook

#filename = "gradedatamissing.csv" # uncomment this in Jupyter Notebook

df_missing = pd.read_csv(filenamprint(pd.__version__)e)

# df.head()

# total missing values

df_missing.isnull().sum()

# show rows with missing values

missing = df_missing['exercise'].isnull()

# missing will only show True/False values

df_missing.loc[missing]

names = ['Jessica','John','Bob','Jessica','Mary','John','Mel','Mel']

grades = [95,78,76,95,77,78,99,100]

GradeList = list(zip(names,grades))

df = pd.DataFrame(data = GradeList, columns=['Names', 'Grades'])

df

# boolean values for if there's another row with the exact values in each column

dupe = df.duplicated()

# duplicate of Jessica, 95; John, 78

# returns false on first instance of duplicate row

dupe

df.loc[dupe]

colA = [10, 20, 40, 50]

colB = ['no', 'yes', 'yes', 'no']

A_B = list(zip(colA, colB))

df_A = pd.DataFrame(data=A_B, columns=['A', 'B'])

df_A

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

colA = [10, 20, 40, 50]

colB = ['no', 'yes', 'yes', 'no']

df_B = pd.DataFrame(data=A_B, columns=['A', 'B'], index=index)

df_B

# loc: label based indexing

df_A.loc[0]

# iloc: select row by index(position) number

df_A.iloc[0]

# loc: label based indexing

df_B.loc['b']

df_B.iloc[1]

# let's see how it gets tricky

df_C = df_A.copy()

df_C

index = [2,1,0,3]

colA = [10, 20, 40, 50]

colB = ['no', 'yes', 'yes', 'no']

df_D = pd.DataFrame(data=A_B, columns=['A', 'B'], index=index)

df_D

# works like python slicing; give me index 0 through the index before 3

df_C.iloc[0:3]

# give me the index label '0' and everything in between and include the label '3'

df_C.loc[0:3]

# gave index position 0 through the index position before 3

df_D.iloc[0:3]

# gave the label '0' and all the rows that are in between and ending with the row with label '3'

df_D.loc[0:3]

# import glob # uncomment this in Jupyter Notebook

data_files = [data1.xlsx, data2.xlsx, data3.xlsx] # comment out this line in Jupyter Notebook

all_data = pd.DataFrame()

for f in data_files: # comment out this line in Jupyter Notebook 

# for f in glob.glob("data*.xlsx"): # uncomment this in Jupyter Notebook

    df = pd.read_excel(f)

    all_data = all_data.append(df, ignore_index=True)

all_data.count() # each data file had 100 rows

from numpy import random

names = ['Bob','Jessica','Mary','John','Mel']

random.seed(500)

random_names = [names[random.randint(low=0,high=len(names))] 

 				for i in range(1000)]

births = [random.randint(low=0,high=1000) 

 		  for i in range(1000)]

BabyDataSet = list(zip(random_names,births))

df = pd.DataFrame(data = BabyDataSet, columns=['Names', 'Births'])

df.head()