/kaggle/input/students-exam-scores/Original_data_with_more_rows.csv/kaggle/input/students-exam-scores/Expanded_data_with_more_features.csv

In [2]:

import matplotlib as pltimport seaborn as sns

In [3]:

df= pd.read_csv('/kaggle/input/students-exam-scores/Original_data_with_more_rows.csv')

In [6]:

df.isnull().sum()

Out[6]:

Unnamed: 0 0Gender 0EthnicGroup 0ParentEduc 0LunchType 0TestPrep 0MathScore 0ReadingScore 0WritingScore 0dtype: int64

In [7]:

df.rename(columns={'Unnamed: 0':'Id'},inplace=True)

In [8]:

pv1=pd.pivot_table(df,index='Gender',values=['MathScore','ReadingScore','WritingScore'])

In [10]:

plt.style.use('ggplot')

In [11]:

p1=pv1.plot(kind='barh',y=['MathScore','ReadingScore','WritingScore'],edgecolor='black',linewidth=2,figsize=(12,8),title='Males vs Females ')p1.bar_label(p1.containers[0], label_type='edge',padding=0.5,bbox={"boxstyle": "round", "pad": 0.6, "facecolor": "white", "edgecolor": "#1c1c1c", "linewidth" : 2.5, "alpha": 1})p1.bar_label(p1.containers[1], label_type='edge',padding=0.5,bbox={"boxstyle": "round", "pad": 0.6, "facecolor": "white", "edgecolor": "#1c1c1c", "linewidth" : 2.5, "alpha": 1})p1.bar_label(p1.containers[2], label_type='edge',padding=0.5,bbox={"boxstyle": "round", "pad": 0.6, "facecolor": "white", "edgecolor": "#1c1c1c", "linewidth" : 2.5, "alpha": 1})

Out[11]:

[Text(0.5, 0, '72.8977'), Text(0.5, 0, '63.9787')]

From the Bar chart above, it's clear that males are better in Math while females are batter in writing.

In [12]:

df['OverallScore']=(df['MathScore']+df['ReadingScore']+df['WritingScore'])/3

That's better!

In [14]:

pv2=pd.pivot_table(df,index='EthnicGroup',values=['OverallScore'])p2=pv2.plot(kind='bar',y=['OverallScore'],edgecolor='black',linewidth=2,figsize=(12,7),title='Overallscore by EthnicGroup ')p2.bar_label(p2.containers[0], label_type='edge',padding=0.5,bbox={"boxstyle": "round", "pad": 0.6, "facecolor": "white", "edgecolor": "#1c1c1c", "linewidth" : 2.5, "alpha": 1})

Out[14]:

[Text(0, 0.5, '65.0971'), Text(0, 0.5, '65.7512'), Text(0, 0.5, '66.855'), Text(0, 0.5, '69.7398'), Text(0, 0.5, '74.2322')]

In [15]:

pv3=pd.pivot_table(df,index='Gender',values=['OverallScore'])p3=pv3.plot(kind='barh',y=['OverallScore'],edgecolor='black',linewidth=2,figsize=(12,7),title='Overallscore by Gender ')p3.bar_label(p3.containers[0], label_type='edge',padding=0.5,bbox={"boxstyle": "round", "pad": 0.6, "facecolor": "white", "edgecolor": "#1c1c1c", "linewidth" : 2.5, "alpha": 1})

Out[15]:

[Text(0.5, 0, '70.0848'), Text(0.5, 0, '66.4524')]

In [16]:

pv4=pd.pivot_table(df,index='ParentEduc',values=['OverallScore'])pv4=pv4.sort_values(by=['OverallScore'],ascending=True)p4=pv4.plot(kind='barh',y=['OverallScore'],edgecolor='black',linewidth=2,figsize=(12,7),title='Overallscore by Parent Education level ')p4.bar_label(p4.containers[0], label_type='edge',padding=0.5,bbox={"boxstyle": "round", "pad": 0.6, "facecolor": "white", "edgecolor": "#1c1c1c", "linewidth" : 2.5, "alpha": 1})

Out[16]:

[Text(0.5, 0, '64.1895'), Text(0.5, 0, '65.9195'), Text(0.5, 0, '68.2617'), Text(0.5, 0, '70.0903'), Text(0.5, 0, '72.4781'), Text(0.5, 0, '74.2054')]

In [17]:

p4=pv4.plot(kind='line',y=['OverallScore'],linewidth=4,figsize=(12,7),title='Overallscore by Parent Education level ')

In [18]:

pv6=pd.pivot_table(df,index='LunchType',values=['OverallScore'])pv6=pv6.sort_values(by=['OverallScore'],ascending=True)p6=pv6.plot(kind='barh',y=['OverallScore'],edgecolor='black',linewidth=2,figsize=(12,7),title='Overallscore by Lunch Type ')p6.bar_label(p6.containers[0], label_type='edge',padding=0.5,bbox={"boxstyle": "round", "pad": 0.6, "facecolor": "white", "edgecolor": "#1c1c1c", "linewidth" : 2.5, "alpha": 1})

Out[18]:

[Text(0.5, 0, '62.0637'), Text(0.5, 0, '71.6342')]

In [19]:

pv5=pd.pivot_table(df,index='TestPrep',values=['OverallScore'])pv5=pv5.sort_values(by=['OverallScore'],ascending=True)

p5=pv5.plot(kind='barh',y=['OverallScore'],edgecolor='black',linewidth=2,figsize=(12,7),title='Overallscore by Test Preparation ')

p5.bar_label(p5.containers[0], label_type='edge',padding=0.5,bbox={"boxstyle": "round", "pad": 0.6, "facecolor": "white", "edgecolor": "#1c1c1c", "linewidth" : 2.5, "alpha": 1})

Out[19]:

[Text(0.5, 0, '65.8889'), Text(0.5, 0, '72.821')]

In [20]:

df['OverallScore'].plot(kind='hist',bins=40,figsize=(12,7),edgecolor='black',title='Overallscore distribution')

Out[20]:

<AxesSubplot:title={'center':'Overallscore distribution'}, ylabel='Frequency'>

Most of the exam score range from 60 to 80!

In [21]:

df2=df.copy()

In [22]:

df2=df2.drop(columns='Id',axis=1)

In [24]:

df2['Gender']=pd.factorize(df2.Gender)[0]df2['EthnicGroup']=pd.factorize(df2.EthnicGroup)[0]

df2['ParentEduc']=pd.factorize(df2.ParentEduc)[0]

df2['LunchType']=pd.factorize(df2.LunchType)[0]df2['TestPrep']=pd.factorize(df2.TestPrep)[0]

In [25]:

sns.heatmap(df2.corr(),annot = True, cmap= 'YlGnBu', fmt= '.2f')

Out[25]:

<AxesSubplot:>

From the heatmap above, it's obevious that Lunch type and Test preparation has a higher impact on the student's overallscore. Gender, Ethnic group and Parent education level has an impact also on the score.

In [26]:

sns.pairplot(df)

Out[26]:

<seaborn.axisgrid.PairGrid at 0x70179a6a0650>

Original notebook: