DEV Community: muradalhasan

Seaborn-High Level Graphical Representation

muradalhasan — Tue, 15 Aug 2023 09:54:14 +0000

Data visualization library based on Matplotlib. Provides High Level Graphical Interface.

import seaborn as sns
import matplotlib.pyplot as plt

loading dataset,# draw lineplot,# setting the x limit of the plot# # changing the theme to dark

data = sns.load_dataset("iris")
sns.lineplot(x="sepal_length", y="sepal_width", data=data)
plt.xlim(5)
sns.set_style("dark")
plt.show()

figure() method.

data = sns.load_dataset("iris")
plt.figure(figsize = (4, 4))
sns.lineplot(x="sepal_length", y="sepal_width", data=data)
sns.despine()
plt.show()

set_context() method

Syntax:

set_context(context=None, font_scale=1, rc=None)

data = sns.load_dataset("iris")
sns.lineplot(x="sepal_length", y="sepal_width", data=data)
sns.set_context("poster")
plt.show()

*Subplot Already Discued in Matplotlib

Relational Plots-->Relational plots are used for visualizing the statistical relationship between the data points

Syntax:

seaborn.relplot(x=None, y=None, data=None, **kwargs)

data = sns.load_dataset("iris")
sns.relplot(x='sepal_width', y='species', data=data)

plt.show()

Scatter Plot

Syntax:

seaborn.scatterplot(x=None, y=None, data=None, **kwargs)

data = sns.load_dataset("iris")
sns.scatterplot(x='sepal_length', y='sepal_width', data=data)
plt.show()

Line Plot

Syntax:

seaborn.lineplot(x=None, y=None, data=None, **kwargs)

data = sns.load_dataset("iris")
sns.lineplot(x='sepal_length', y='species', data=data)
plt.show()

Bar Plot

Syntax:

barplot([x, y, hue, data, order, hue_order, …])

data = sns.load_dataset("iris")
sns.barplot(x='species', y='sepal_length', data=data)
plt.show()

Count Plot

Syntax:

countplot([x, y, hue, data, order, …])

data = sns.load_dataset("iris")

sns.countplot(x='species', data=data)
plt.show()

Box Plot

Syntax:

# boxplot([x, y, hue, data, order, hue_order, …])

data = sns.load_dataset("iris")

sns.boxplot(x='species', y='sepal_width', data=data)
plt.show()

Violinplot

Syntax:

violinplot([x, y, hue, data, order, …]

data = sns.load_dataset("iris")

sns.violinplot(x='species', y='sepal_width', data=data)
plt.show()

Stripplot

Syntax:

stripplot([x, y, hue, data, order, …])

data = sns.load_dataset("iris")

sns.stripplot(x='species', y='sepal_width', data=data)
plt.show()

Swarmplot

Syntax:

swarmplot([x, y, hue, data, order, …])

data = sns.load_dataset("iris")

sns.swarmplot(x='species', y='sepal_width', data=data)
plt.show()

Histogram

Syntax:

histplot(data=None, *, x=None, y=None, hue=None, **kwargs)

data = sns.load_dataset("iris")

sns.histplot(x='species', y='sepal_width', data=data)
plt.show()

Distplot

Syntax:

distplot(a[, bins, hist, kde, rug, fit, …])

data = sns.load_dataset("iris")

sns.distplot(data['sepal_width'])
plt.show()

Jointplot

Syntax:

jointplot(x, y[, data, kind, stat_func, …])

data = sns.load_dataset("iris")
sns.jointplot(x='species', y='sepal_width', data=data)
plt.show()

Pairplot

Syntax:

pairplot(data[, hue, hue_order, palette, …])

data = sns.load_dataset("iris")
sns.pairplot(data=data, hue='species')
plt.show()

Rugplot

Syntax:

rugplot(a[, height, axis, ax])

data = sns.load_dataset("iris")
sns.rugplot(data=data)
plt.show()

KDE Plot

Syntax:

seaborn.kdeplot(x=None, *, y=None, vertical=False, palette=None, **kwargs)

data = sns.load_dataset("iris")
sns.kdeplot(x='sepal_length', y='sepal_width', data=data)
plt.show()

Regression Plots

Syntax:

seaborn.lmplot(x, y, data, hue=None, col=None, row=None, **kwargs)

data = sns.load_dataset("tips")
sns.lmplot(x='total_bill', y='tip', data=data)
plt.show()

Regplot

Syntax:

seaborn.regplot( x, y, data=None, x_estimator=None, **kwargs)

data = sns.load_dataset("tips")
sns.regplot(x='total_bill', y='tip', data=data)
plt.show()

Heatmap

Syntax:

seaborn.heatmap(data, *, vmin=None, vmax=None, cmap=None, center=None, annot_kws=None, linewidths=0, linecolor=’white’, cbar=True, **kwargs)

data = sns.load_dataset("tips")
# correlation between the different parameters
tc = data.corr()
sns.heatmap(tc)
plt.show()

Matplotlib

muradalhasan — Wed, 02 Aug 2023 13:54:08 +0000

Matplotlib is a low level graph plotting and open source library was created by John D. Hunter.There are two types of method to creat a matplotlib plot functional and Object oriented method.
*To plot a function we need two array of ponits and plot() function then show() function

xpoints = [0, 6]
ypoints = [0, 120]

plt.plot(xpoints, ypoints)
plt.show()

and it will plot a graph according to x-ponits and y-pointes

size(ms),marker_color(mec),ful_marker_color(mec),linestyle, linewidth teg of plot function

ypoints = np.array([3, 8, 1, 10])

plt.plot(ypoints, marker = 'o',color="hotpink", ms = 20,mec="yellow",mfc="blue",linestyle="--",linewidth=10)
plt.show()

marker function mark the satisfied points
color function gives color of lines
ms decide the size
mec decide the marker circle color
mfc decide marker color
linestyle decides linestyle
linewidth decides linewidth of line
*lebeling the axis and plot title

x = np.array([80, 85, 90, 95, 100, 105, 110, 115, 120, 125])
y = np.array([240, 250, 260, 270, 280, 290, 300, 310, 320, 330])

plt.plot(x, y)
plt.title("This is an example title",color='red',size=20,loc='right')
plt.xlabel("Average Pulse",color='blue',size=20,family='serif')
plt.ylabel("Calorie Burnage",color='yellow',size=20,family='serif')

plt.show()

*grid() functionn and it's attributes

plt.grid(axis='x')
plt.show()

plt.grid(color = 'green', linestyle = '--', linewidth = 0.5)

*subplot function

plt.subplot(2, 3, 1)
plt.plot(x,y)

2 in subplot means 2 rows and 3 column and position of that subplot is 1
*scatter diagram.we can plot multiple scatter in a single graph

x = np.array([5,7,8,7,2,17,2,9,4,11,12,9,6])
y = np.array([99,86,87,88,111,86,103,87,94,78,77,85,86])
plt.scatter(x, y)
plt.scatter(x, y, color = 'hotpink')

#day two, the age and speed of 15 cars:
x = np.array([2,2,8,1,15,8,12,9,7,3,11,4,7,14,12])
y = np.array([100,105,84,105,90,99,90,95,94,100,79,112,91,80,85])
plt.scatter(x, y)
plt.scatter(x, y, color = '#88c999')
plt.show()

*sctter diagram attributes

x = np.array([5,7,8,7,2,17,2,9,4,11,12,9,6])
y = np.array([99,86,87,88,111,86,103,87,94,78,77,85,86])
colors = np.array([0, 10, 20, 30, 40, 45, 50, 55, 60, 70, 80, 90, 100])
sizes = np.array([20,50,100,200,500,1000,60,90,10,300,600,800,75])

plt.scatter(x, y, c=colors, s=sizes, alpha=0.5, cmap='nipy_spectral')

plt.colorbar()

plt.show()

*Bar diagram

x = np.array(["A", "B", "C", "D"])
y = np.array([3, 8, 1, 10])

plt.bar(x,y)
plt.show()

*#to draw horiozontal bar

x = np.array(["A", "B", "C", "D"])
y = np.array([3, 8, 1, 10])

plt.barh(x, y)
plt.show()

*plt.bar() attribues

x = np.array(["A", "B", "C", "D"])
y = np.array([3, 8, 1, 10])
colors=np.array(["red","black","hotpink","yellow"])#we can add single color also customise color for every bar

plt.bar(x, y, color = colors)
plt.show()

width = 0.1 bar takes width size like color we can make all bars same size or customise size like color

*A histogram:

x = np.random.normal(170, 10, 250)

plt.hist(x,color="hotpink")
plt.show()

*pie chart and it's attributes

y = np.array([15, 10, 25, 50])

plt.pie(y)
plt.show()

The label parameter must be an array with one label for each wedge

y = np.array([35, 25, 25, 15])
mylabels = ["Apples", "Bananas", "Cherries", "Dates"]

plt.pie(y, labels = mylabels)
plt.show()

Pull the "Apples" wedge 0.2 from the center of the pie:

y = np.array([35, 25, 25, 15])
mylabels = ["Apples", "Bananas", "Cherries", "Dates"]
myexplode = [0.2, 0, 0, 0]

plt.pie(y, labels = mylabels, explode = myexplode)
plt.show()
# Add a shadow to the pie chart by setting the shadows parameter to True:
plt.pie(y, labels = mylabels, explode = myexplode, shadow = True)
plt.show()

Specify a new color for each wedge:

y = np.array([35, 25, 25, 15])
mylabels = ["Apples", "Bananas", "Cherries", "Dates"]
mycolors = ["black", "hotpink", "b", "#4CAF50"]

plt.pie(y, labels = mylabels, colors = mycolors)
plt.show()

To add a list of explanation for each wedge, use the legend() function:

y = np.array([35, 25, 25, 15])
mylabels = ["Apples", "Bananas", "Cherries", "Dates"]

plt.pie(y, labels = mylabels)
plt.legend()
# To add a header to the legend, add the title parameter to the legend function. 
plt.legend(title = "Four Fruits:")
plt.show()

Pandas-Data Manipulation and Analysis

muradalhasan — Sat, 21 Jan 2023 18:51:40 +0000

** Introduction to Pandas**

Pandas is basically a Python library that is used for analyzing, cleaning, exploring, and manipulating data. Pandas allow us to analyze big data and make conclusions based on statistical theories. Pandas can clean messy data sets, and make them readable and relevant. Relevant data is very important in data science.

Installation and importing Pandas.

First of all, open your jupyter notebook, then go to your command panel and write “pip install Pandas”. Here we go, you are done with the installation of the Pandas library. Now, to use the Pandas library you need to import Pandas in your notebook, to do so write “import Pandas as PD”. You are ready to use Pandas Library.
Finally, check if the installation was successful.

import pandas as pd
print(pd.version)

Getting Started with Pandas

Series

A series can actually hold pretty much almost all types of data
objects with python as its data point. Series also can take lots of parameters (data, index, dtype, name, copy, fastpath). To create a pandas series we just need a list initially it will generate random index we can also manipulate index by sending a list of index with the index parameter.

Just look at the picture above. we have used an integer list and series automatically generated index but in the 2nd and 3rd series we have customized the index as we wanted that's the beauty of Pandas.

We can also use a dictionary and Pandas series will use its key as index .

Use the index option to specify only the words you wish to be included in the Series, leaving out the rest of the words in the dictionary. Two series can added.

Data Frames
Data Frame is known as the main tool of Pandas. It takes multiple parameters like data, column, rows, dtype, copy. To start with Data Frames we have some necessary setting in our Jypyter. As we don't have any specific data so, we will use random data to learn Data Frames.

import numpy as np
import pandas as pd
from numpy.random import randn
np.random.seed(101)
df=pd.DataFrame(randn(5,4),["A","B","C","D","E"],["X","Y","Z","P"])
df

output


       X           Y           Z          P
A   2.706850    0.628133    0.907969    0.503826
B   0.651118    -0.319318   -0.848077   0.605965
C   -2.018168   0.740122    0.528813    -0.589001
D   0.188695    -0.758872   -0.933237   0.955057
E   0.190794    1.978757    2.605967    0.683509

Data Frames is just bunches of series. we can access those series by following command.

 **df["X"]
output
    X          Y
A   2.706850    0.628133
B   0.651118    -0.319318
C   -2.018168   0.740122
D   0.188695    -0.758872
E   0.190794    1.978757

more than one series is a data frames

**df[["X","Y"]]
        X         Y
A   2.706850    0.628133
B   0.651118    -0.319318
C   -2.018168   0.740122
D   0.188695    -0.758872
E   0.190794    1.978757

we can also add or drop any column.
To add any column.

df["updated"]=df["X"]+df["Y"]
df["updated"]

A    3.334983
B    0.331800
C   -1.278046
D   -0.570177
E    2.169552

it will add the number column x and y and create a column which will be added as updated.
To drop any column.

df.drop("updated",axis=1,inplace=True)
df
         X          Y          Z        p
A   2.706850    0.628133    0.907969    0.503826    
B   0.651118    -0.319318   -0.848077   0.605965    
C   -2.018168   0.740122    0.528813        -1.278046
D   0.188695    -0.758872   -0.933237   0.955057    
E   0.190794    1.978757    2.605967    0.683509

as we have set inplace as true so the updated column will drop permanently.
we can also add or drop any rows or columns by using this command.
df.drop("D")
we just have to put the row or column name and it will be droped automatically.
we can check dimension of a data frames by
df.shape

Conditional selection in Data Frames

df>0
output
      X     Y       Z        P      updated
A   True    True    True    True    True
B   True    False   False   True    True
C   False   True    True    False   False
D   True    False   False   True    False
E   True    True    True    True    True

we can also specify only false value

booldf=df>0
df[booldf]
output
       X            Y           Z            P  
A   2.706850    0.628133    0.907969    0.503826    
B   0.651118    NaN          NaN            0.605965
C   NaN         0.740122    0.528813    NaN
D   0.188695    NaN          NaN           0.955057 
E   0.190794    1.978757    2.60596     2.169552

we can also use specific rows or column or both

df["X"]>0

starting with multiple conditions

df[(df["X"]>0) | (df["Y"]>1)]

** skipping index reset it's bit harder to explain index reset in that blog.

Pipe Index Oparetor

index level

a="G1 G1 G1 G2 G2 G2".split()
outside=a
inside=[1,2,3,1,2,3]
higher_index=list(zip(outside,inside))
higher_index=pd.MultiIndex.from_tuples(higher_index)

now we can use that pipe index to built a multi level index

df=pd.DataFrame(randn(6,2),higher_index,["A","B"])

to get sub data frames from a multilevel index

df.loc["G1"]
df.loc["G1"].loc[1]

Missing Data

d={"A":[1,2,np.nan],"B":[1,np.nan,np.nan],"C":[1,2,3]}
df=pd.DataFrame(d)
output
       A    B   C
0   1.0 1.0 1
1   2.0 NaN 2
2   NaN NaN 3

to drop any row that hold missing data

df.dropna()
it will drop that rows that has one or more NaN value 
df.dropna(axis=1)
it will drop that coloum that has one or more NaN value

to fill any missing data

df.fillna(value="FIll  VALUE")

to chnage specific value

df["A"].fillna(value=3)

mean function

df["A"].fillna(value=df["A"].mean())

GROUP BY

all functions of group by

data={"Company":["GOOG","GOOG","ghgh","fdff"],"Persoon":["Mridul","Mohosin","Sabid","Raihan"],"Sales":[10,12,34,45]}
df=pd.DataFrame(data)
df
bycomp=df.groupby("Company")
to find minimun valuee
bycomp.mean()

total sell of a company

bycomp.sum()

standered division

bycomp.std()

to see sells of a specific company

bycomp.sum().loc["GOOG"]

df.groupby("Company").sum().loc["GOOG"]
df.groupby("Company").count()
df.groupby("Company").max()
df.groupby("Company").min()
df.groupby("Company").describe()
df.groupby("Company").describe().transpose()
df.groupby("Company").describe().transpose()["GOOG"]

merging,joining,concatenating

dic={
    'col1':[1,2,3,4,5],
    "col2":[45,5,77,88,5],
    "col3":["abc","def","ghi","jkl","mno"],
}
df=pd.DataFrame(dic)
df.head()

to find unique values in a specific coloum

df["col2"].unique()

it will return a arrays with unique values

we can aslo check the lenght of that array by len function

len(df["col2"].unique())

nunique is an alternatives of len function

df["col2"].nunique()

unique value count method

df["col2"].value_counts()

filtering data using conditional statement

df[df["col1"]>2]

pandas can use customizer function

Thank you

Introduction To NumPy

muradalhasan — Fri, 18 Nov 2022 17:56:57 +0000

Introduction To Numpy

NumPy, which stands for Numerical Python, is a library consisting of multidimensional array objects and a collection of functions for processing those arrays.Basically NumPy arrays are 50x faster then python lists that’s why we may have to use NumPy to handle big data .The array object in NumPy is called “ndarray”, it provides a lot of supporting functions that make working with “ndarray” very easy.Arrays are very frequently used in data science, where speed and resources are very important.

INSTALLATION AND IMPORTING NumPy

First of all open your Jupyter notebook,then go to your command panel write “pip install numpy”.Here we go,you are done with installation of NumPy library .Now,to use NumPy library you need to import NumPy in your notebook,to do so write “import numpy as np”.You are ready to use NumPy Library.

GETTING STARTED WITH NumPy
1.Converting a list into a array
Let say,
Import numpy as np
my_list=[1,2,3,4,5]

#one dimension array

        arr=np.array(my_list)
        print(arr)  #it will print a array
        print(type(arr)) ##type array
        ######Two Dimensional array
        my_list=[[1,2,3][4,5,6][6,7,8]]
        arr=np.array(my_list)
        ##indexing of arrays are  similar to list indexing

2.Creating automatic one dimensional array
arr=np.arange(a,b,c) #a=star(inclusive),b=end(exclusive),c=step
##it will generate array like for loop
**Split of an array

import numpy as np

arr = np.array([[1, 2, 3], [4, 5, 6], [7, 8, 9], [10, 11, 12], [13, 14, 15], [16, 17, 18]])

newarr = np.array_split(arr, 3, axis=1)

print(newarr)

**output

[array([[ 1],
       [ 4],
       [ 7],
       [10],
       [13],
       [16]]), array([[ 2],
       [ 5],
       [ 8],
       [11],
       [14],
       [17]]), array([[ 3],
       [ 6],
       [ 9],
       [12],
       [15],
       [18]])]

**to find a specific eliment index

import numpy as np

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

x = np.where(arr == 4)

print(x)

output

(array([3, 5, 6], dtype=int64),)

*to sort an array

arr = np.array([3, 2, 0, 1])

print(np.sort(arr))

*output

[0 1 2 3]

*to filter an array with some specific condition

arr = np.array([41, 42, 43, 44])
filter_arr = []
for element in arr:
  if element > 42:
    filter_arr.append(True)
  else:
    filter_arr.append(False)
newarr = arr[filter_arr]
print(filter_arr)
print(newarr)

*output

[False, False, True, True]
[43 44]

3.Generating matrix with Natural Numbers
arr=np.zeros(a) #a → length
output=array([0., 0., 0.]) #a=3
arr=np.zeros((a,b)) #a→row,by→column
output=array([[0., 0., 0.], #here a=2 and b=3
[0., 0., 0.]])

   ####we can also generate with other numbers
   arr=np.ones((2,3))
   arr=np.twos((5,5))

4.Generating fractional array (linspace)
arr=np.linspace(x,y,z)

x →starting number(inclusive),y →ending number(inclusive),z →total number of numbers

5.Generating Identity metrics
arr=np.eye(x) ##x →Number of rows and column
6.the random number numpy genarate is basically psudo random number cause there is and algorithm that generate random number random can generate three types of random number interger,float and negative.
arrays of random numbers
arr=np.random.rand(a) #a —>single line length
arr=np.random.rand(a,b) #a by b matrix
np.random.randn(2,4) #row column dimension float
np.random.randint(1,100) #it will generate a random number from 0 to 99
np.random.randint(1,100,10) # it will generate 10 random int number from 1 to 100 and 100 exclusive
7.#reshape method
arr= np.arange(25)
arr.reshape(5,5)
8.#to return max value of a array
arr.max()
9.# to return min
arr.min()
10.# to see max and min value location # arg
arr.argmax()
arr.argmin()
11.##numpy indexing and selection
arr=np.arange(1,10)

arr

arr[8]
arr[2:5]
12.# to change the value by using index
arr[0:3]=100
arr
13.##to access any element from the arr_2D
arr_2D=np.array([[1,2,3],[4,5,6],[7,8,9]])
arr_2D[0][1] #matrix concept
14.##conditional selection
arr=np.arange(1,15)
arr
arr>5
bool_arr=arr>5
arr[bool_arr]
15.###numpy operations
import numpy as np
arr=np.arange(10)
Arr+arr
Arr+100

16.#numpy subtract
arr-arr
17.##numpy multiplication
arr*arr
19.#numpy division
arr/arr

0 by 0 infinity

20.##numpy exponential
arr**2
np.sin(arr)