In [1]:
import pandas as pd
import seaborn as sns
import numpy as np
from matplotlib import pyplot as plt
In [3]:
data = pd.read_csv("scholarship.csv")
In [4]:
data.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1000 entries, 0 to 999
Data columns (total 3 columns):
 #   Column                  Non-Null Count  Dtype  
---  ------                  --------------  -----  
 0   semester_percentage     1000 non-null   float64
 1   scholarship_exam_marks  1000 non-null   int64  
 2   got_scholarship         1000 non-null   int64  
dtypes: float64(1), int64(2)
memory usage: 23.6 KB
In [5]:
data.head(5)
Out[5]:
semester_percentage scholarship_exam_marks got_scholarship
0 71.9 26 1
1 74.6 38 1
2 75.4 40 1
3 64.2 8 1
4 72.3 17 0
In [23]:
plt.subplot(1,2,1)
sns.histplot(data["semester_percentage"],  kde=True, stat="density", linewidth=0)
# sns.distplot(data["semester_percentage"])
plt.subplot(1,2,2)
sns.histplot(data["scholarship_exam_marks"],  kde=True, stat="density", linewidth=0)
plt.show()

# Hence we can Observe that the semester_percentage is a Normal Distribution and the scholarship_exam_marks is a Skewed Distribution
No description has been provided for this image
In [ ]:
print("Mean = ", data["semester_percentage"].mean())
print("Std = " ,data["semester_percentage"].std())
print("Max = " ,data["semester_percentage"].max())
print("Min = " ,data["semester_percentage"].min())

Mean =  69.6124
Std =  6.158978751323897
Max =  91.2
Min =  48.9
In [9]:
#Finding the boundary values
# mean = mue , std = sigma , z = (x-mue)/sigma normal Distribution

Highest_allowed = data["semester_percentage"].mean() + 3*data["semester_percentage"].std()
Lowest_allowed = data["semester_percentage"].mean() - 3*data["semester_percentage"].std()
print("Highest_allowed : " , Highest_allowed)
print("Lowest_allowed : " , Lowest_allowed)

Highest_allowed :  88.08933625397168
Lowest_allowed :  51.13546374602831
In [10]:
# finding the outliers
data[(data["semester_percentage"]<Lowest_allowed) | (data["semester_percentage"]>Highest_allowed) ]
Out[10]:
semester_percentage scholarship_exam_marks got_scholarship
485 49.2 44 1
995 88.7 44 1
996 91.2 65 1
997 48.9 34 0
999 49.0 10 1
In [11]:
#Trimming
data[(data["semester_percentage"]>Lowest_allowed) & (data["semester_percentage"]<Highest_allowed) ]
Out[11]:
semester_percentage scholarship_exam_marks got_scholarship
0 71.9 26 1
1 74.6 38 1
2 75.4 40 1
3 64.2 8 1
4 72.3 17 0
... ... ... ...
991 70.4 57 0
992 62.6 12 0
993 67.3 21 1
994 64.8 63 0
998 86.2 46 1

995 rows × 3 columns

In [ ]:
# another method to trim
data["z_score"] = (data["semester_percentage"]-data["semester_percentage"].mean())/data["semester_percentage"].std()
In [ ]:
data[data["z_score"]<-3]
Out[ ]:
semester_percentage scholarship_exam_marks got_scholarship z_score
485 49.2 44 1 -3.314251
997 48.9 34 0 -3.362960
999 49.0 10 1 -3.346724
In [ ]:
data[data["z_score"]>3]
Out[ ]:
semester_percentage scholarship_exam_marks got_scholarship z_score
995 88.7 44 1 3.099150
996 91.2 65 1 3.505062
In [ ]:
data[(data["z_score"]>-3) & (data["z_score"]<3)]
Out[ ]:
semester_percentage scholarship_exam_marks got_scholarship z_score
0 71.9 26 1 0.371425
1 74.6 38 1 0.809810
2 75.4 40 1 0.939701
3 64.2 8 1 -0.878782
4 72.3 17 0 0.436371
... ... ... ... ...
991 70.4 57 0 0.127878
992 62.6 12 0 -1.138565
993 67.3 21 1 -0.375452
994 64.8 63 0 -0.781363
998 86.2 46 1 2.693239

995 rows × 4 columns

In [14]:
#capping
upper_limit = data["semester_percentage"].mean()+ 3* data["semester_percentage"].std()
lower_limit = data["semester_percentage"].mean()- 3* data["semester_percentage"].std()
In [15]:
data["semester_percentage"] =np.where(data["semester_percentage"]>upper_limit,
        upper_limit,
        np.where(data["semester_percentage"]<lower_limit, lower_limit, data["semester_percentage"] ))
In [16]:
data
Out[16]:
semester_percentage scholarship_exam_marks got_scholarship
0 71.900000 26 1
1 74.600000 38 1
2 75.400000 40 1
3 64.200000 8 1
4 72.300000 17 0
... ... ... ...
995 88.089336 44 1
996 88.089336 65 1
997 51.135464 34 0
998 86.200000 46 1
999 51.135464 10 1

1000 rows × 3 columns

In [17]:
# Plotting of Normal Distribution
from scipy import stats
import numpy as np
from matplotlib import pyplot as pyt
In [18]:
x_data = np.arange(-5,5,0.00001)
y_data = stats.norm.pdf(x_data,0,1)
z_data = stats.norm.pdf(x_data , 0 ,2)
In [19]:
plt.plot(x_data , y_data)
Out[19]:
[<matplotlib.lines.Line2D at 0x2eb38372610>]
No description has been provided for this image
In [20]:
plt.plot(x_data,z_data)
Out[20]:
[<matplotlib.lines.Line2D at 0x2eb3c6f7e10>]
No description has been provided for this image
In [ ]: