Upload 20 files

app.py

@@ -0,0 +1,84 @@
+import streamlit as st
+from module.cards import welcome, about_me
+from module.dataset import eda
+from module.load_data import data
+from module.data_viz import visualization
+import numpy as np
+from datetime import datetime
+
+global LOW, HIGH, SEQUENCE, RESET
+
+LOW = 10.0*np.ones((46,7))
+HIGH = 20.0*np.ones((46,7))
+SEQUENCE = ['seq 1','seq 2','seq 3','seq 4','seq 5','seq 6','average_marks']
+RESET = True
+
+st.set_page_config(
+    page_title="ICT Mark's Dashboard",
+    page_icon="📢",
+    layout="wide")
+
+
+
+st.title("📊 :red[ICT Mark's] Dashboard")
+
+welcome()
+about_me()
+
+date = f"{datetime.today().date()} {datetime.today().hour}:{datetime.today().minute}:{datetime.today().second}"
+
+#page = st.selectbox('Select page:', ['Dataset','EDA', 'Data Viz'])
+tab1, tab2, tab3 = st.tabs(['**Dataset**','**EDA**', '**Data Viz**'])
+
+with  tab1:
+    
+    st.info(f"You select Dataset page at the {date}", icon="💂")
+    col1, col2, col3, col4 = st.columns(4)
+
+    if col1.button(':orange[**highlight >10/20**]'):
+        st.dataframe(data.style.highlight_between(axis=None, color="#fffd75", left=LOW, right=HIGH,
+        subset=SEQUENCE), use_container_width=True)
+        RESET = False
+
+    if col2.button(':orange[**highlight max**]'):
+        st.dataframe(data.style.highlight_max(color="green", subset=SEQUENCE), use_container_width=True)
+        RESET = False
+
+    if col3.button(':orange[**highlight min**]'):
+        st.dataframe(data.style.highlight_min(color="red", subset=SEQUENCE), use_container_width=True)
+        RESET = False
+    
+    if RESET:
+        st.dataframe(data, use_container_width=True)
+
+    if col4.button(':black[**reset**]'):
+        st.dataframe(data, use_container_width=True)
+
+
+
+    with st.expander("👁 Read more"):
+        st.markdown("""
+        This dataset is an :red[ICT Mark's] data come from the evaluation of the Tebap students. The attributes are defined as follows:
+
+        > 1. :orange[$seq_{i, i = (1,2,3,4,5,6)}$]: are the 6 sequences that
+        Tebap student's had undergone during school year 2018-2019.
+        > 2. :orange[age]: the age of the students. 
+        > 3. :orange[gender]: the gender of the students (M or F).
+        > 4. :orange[form]: the class of students.
+        > 5. :orange[average_marks]: the annual ICT mark's of each student.
+          """)
+
+with tab2 :
+    st.info(f"You select EDA page at the {date}", icon="💂")
+    st.header('Exploratory Data Analysis')
+    eda()
+	
+
+with tab3:
+    st.info(f"You select Data Viz page at the {date}", icon="💂")
+    st.header('Visualization')
+    visualization() 
+
+
+	
+

images/tree.dot

@@ -0,0 +1,37 @@
+digraph Tree {
+node [shape=box, style="filled, rounded", color="black", fontname="helvetica"] ;
+edge [fontname="helvetica"] ;
+0 [label="seq 6 <= 11.375\nsamples = 46\nvalue = [23, 23]\nclass = Female", fillcolor="#ffffff"] ;
+1 [label="seq 1 <= 1.25\nsamples = 36\nvalue = [13, 23]\nclass = Male", fillcolor="#a9d4f4"] ;
+0 -> 1 [labeldistance=2.5, labelangle=45, headlabel="True"] ;
+2 [label="samples = 4\nvalue = [4, 0]\nclass = Female", fillcolor="#e58139"] ;
+1 -> 2 ;
+3 [label="seq 2 <= 3.125\nsamples = 32\nvalue = [9, 23]\nclass = Male", fillcolor="#86c3ef"] ;
+1 -> 3 ;
+4 [label="seq 4 <= 2.75\nsamples = 5\nvalue = [4, 1]\nclass = Female", fillcolor="#eca06a"] ;
+3 -> 4 ;
+5 [label="samples = 1\nvalue = [0, 1]\nclass = Male", fillcolor="#399de5"] ;
+4 -> 5 ;
+6 [label="samples = 4\nvalue = [4, 0]\nclass = Female", fillcolor="#e58139"] ;
+4 -> 6 ;
+7 [label="seq 1 <= 13.5\nsamples = 27\nvalue = [5, 22]\nclass = Male", fillcolor="#66b3eb"] ;
+3 -> 7 ;
+8 [label="seq 1 <= 4.375\nsamples = 25\nvalue = [3, 22]\nclass = Male", fillcolor="#54aae9"] ;
+7 -> 8 ;
+9 [label="samples = 15\nvalue = [0, 15]\nclass = Male", fillcolor="#399de5"] ;
+8 -> 9 ;
+10 [label="seq 1 <= 5.625\nsamples = 10\nvalue = [3, 7]\nclass = Male", fillcolor="#8ec7f0"] ;
+8 -> 10 ;
+11 [label="samples = 2\nvalue = [2, 0]\nclass = Female", fillcolor="#e58139"] ;
+10 -> 11 ;
+12 [label="seq 6 <= 5.0\nsamples = 8\nvalue = [1, 7]\nclass = Male", fillcolor="#55abe9"] ;
+10 -> 12 ;
+13 [label="samples = 1\nvalue = [1, 0]\nclass = Female", fillcolor="#e58139"] ;
+12 -> 13 ;
+14 [label="samples = 7\nvalue = [0, 7]\nclass = Male", fillcolor="#399de5"] ;
+12 -> 14 ;
+15 [label="samples = 2\nvalue = [2, 0]\nclass = Female", fillcolor="#e58139"] ;
+7 -> 15 ;
+16 [label="samples = 10\nvalue = [10, 0]\nclass = Female", fillcolor="#e58139"] ;
+0 -> 16 [labeldistance=2.5, labelangle=-45, headlabel="False"] ;
+}

module/cards.py

@@ -0,0 +1,75 @@
+import streamlit as st
+from PIL import Image
+
+
+image_file = "images/massock.jpg"
+path_image = "images/SIPE_sketch.jpg"
+image = Image.open(image_file)
+
+text1 = """
+*I am a young Cameroonian named **Massock Batalong Maurice Blaise** with a *Masters in specialty Materials/
+Mechanics option Mechanics at the University of Yaoundé I* and a *Master in Science Mathematics at the
+African Institute for Mathematical Sciences (AIMS-Cameroon)*. And holder of a *Certificate in Big Data
+Analytics with python issued by AIMS-Cameroon*. I am also a Kaggle Notebook & Discussion Expert.*   
+
+- **LinkedIn**: Massock Batalong Maurice Blaise
+- **Email**: lumierebatalong@gmail.com
+"""
+
+text2 = """
+*This dashboard is one little part of my project :green[**Creation of Intelligent system for an Education Performance (ISEP)**].*
+"""
+
+text3 = """
+To construct an intelligent system able to evaluate an educational performance of students of an
+establishment (school, secondary school, training center, university,...) given and to help respectively the
+teacher staff and the administrative staff to arrest well an evolution of school teaching of theirs students.
+Because in future, these students will be human ressources qualified able to releive the challenge and stakes
+of Africa in the domain such that education, health, water, energy, agriculture, infrastructure, etc ...
+"""
+
+
+
+def welcome():
+
+	placeholder = st.empty()
+
+	with placeholder.container():
+		kp1, kp2, kp3, kp4, kp5 = st.columns(5)
+
+		kp1.metric(
+			label= "🗓 **:red[School year]**",
+			value = "2018-2019",
+			delta = ""
+			)
+
+		kp2.metric(
+			label= "👨🎓 :red[**College**]",
+			value = "Tebap",
+			delta = ""
+			)
+		kp3.metric(
+			label="👩🏫 :red[**Form**]",
+			value="6e-5e",
+			delta = ""
+			)
+
+		kp4.metric(
+			label="🏛 :red[**City**]",
+			value="Yaounde",
+			delta = ""
+			)
+		kp5.metric(
+			label="🏴 :red[**Country**] ",
+			value = "Cameroon",
+			delta = "" 
+			)
+
+def about_me():
+	st.sidebar.image(image)
+	st.sidebar.markdown(text1)
+	st.sidebar.markdown(text2)
+	st.sidebar.header("Abstract")
+	st.sidebar.markdown(text3)
+	st.sidebar.image(path_image)
+

module/data_viz.py

@@ -0,0 +1,272 @@
+import streamlit as st 
+import plotly.express as px
+import pandas as pd
+from module.load_data import data   
+from module.dataset import create_data
+import plotly.express as px
+from sklearn.tree import DecisionTreeClassifier, export_graphviz
+import matplotlib.pyplot as  plt
+
+#plt.style.use('seaborn-dark-palette')
+
+def grades(mark):
+	if mark >= 0 and mark < 5:
+		return 'U'
+	elif  mark >= 5 and mark < 9:
+		return 'E'
+	elif mark >= 9 and mark < 12:
+		return 'D'
+	elif mark >= 12 and mark < 15:
+		return 'C'
+	elif mark >= 15 and mark < 18:
+		return 'B'
+	else:
+		return 'A'
+
+def segmentation(data:pd.DataFrame):
+	df = data.copy()
+	factor = pd.factorize(df.form)
+	df['form'] = factor[0]; definition=factor[1]
+
+	cols = [f'seq {i+1}' for i in range(6)] + ['age','form']
+	tree = DecisionTreeClassifier(random_state=31012020)
+	tree.fit(df[cols], df['gender'])
+	
+	export_graphviz(tree, out_file="images/tree.dot", class_names=["Female", "Male"],
+		feature_names=cols, impurity=False, filled=True,rounded=True)
+	result = pd.DataFrame({"Feature":cols, "Gain":tree.feature_importances_})
+	return result
+
+def scatterplot(female=None, male=None, line_separator=None, x_axis=None, y_axis=None):
+	fig = plt.figure(figsize=(5,5))
+	plt.scatter(female[x_axis], female[y_axis], marker='v', color='red', label="Girl", lw=1.2)
+	plt.scatter(male[x_axis], male[y_axis], marker='o', color='blue', label="Boy",lw=1.2)
+	plt.xlabel(x_axis); plt.ylabel(y_axis)
+	plt.title("Decision line", fontweight="bold")
+	for u in  line_separator[0]:
+		plt.vlines(line_separator[0], 0, female[y_axis].max(), color="black",
+		 linestyles="-.", lw=1.75)
+
+	for v in  line_separator[1] :
+		plt.hlines(line_separator[1], -1, female[y_axis].max(), color="black",
+		 linestyles="dashed", lw=1.75)
+
+	plt.legend(loc="best", frameon=True, fancybox=True, shadow=True, title="Gender")
+	return fig
+
+
+df_trim = create_data(data)
+total_students = len(data)
+average_age = data.age.mean()
+gca1 = df_trim["trimester 1"].mean()
+gca2 = df_trim["trimester 2"].mean()
+gca3 = df_trim["trimester 3"].mean()
+annual_gca = (gca1+gca2+gca3)/3
+
+evaluations = ['seq 1', 'seq 2','seq 3','seq 4','seq 5','seq 6']
+
+sgca = pd.DataFrame(columns=['Male','Female'], index=evaluations)
+
+sgca['Male'] = data[data.gender == 'M'][evaluations].mean()
+sgca['Female'] = data[data.gender == 'F'][evaluations].mean()
+
+form_counts = data.form.value_counts()
+gender_counts = data.gender.value_counts()
+
+progression = sgca.diff()
+
+pass_or_fail =pd.DataFrame({seq:data[seq].apply(lambda x: 'Passed' if x >= 10.0 else 'Failed').value_counts().to_dict() for seq in evaluations})   
+
+student_grade =  pd.DataFrame({seq:data[seq].apply(grades).value_counts().to_dict() for seq in evaluations})
+
+##print(student_grade)
+
+bar_polar = px.bar_polar(data_frame=sgca, r='Male', 
+	theta=sgca.index, color='Female', 
+	title='General Class Average (GCA) for each evaluation',
+	barmode="overlay",width=1200, height=500)
+pie_form = px.pie(data_frame=form_counts, names=form_counts.index, values=form_counts,
+ title="Form", hole=0.25, width=500, height=500)
+pie_gender = px.pie(data_frame=gender_counts, names=gender_counts.index, values=gender_counts,
+ title="Gender", hole=0.25, width=500, height=500)
+
+
+feature_importances = segmentation(data)
+
+bar_feature_importance = px.bar(data_frame=feature_importances, x="Feature", y="Gain",
+  title="Feature importances bar", width=500)
+
+with open("images/tree.dot") as f:
+	dot_graph = f.read()
+
+female = data[data.gender == "F"]
+male   = data[data.gender == "M"]
+
+
+def visualization():
+	placeholder1 = st.empty()
+
+	with placeholder1.container():
+		kp1, kp2, kp3, kp4, kp5, kp6 = st.columns(6)
+
+		kp1.metric(
+			label= "**:red[👬 Total students]**",
+			value = total_students,
+			delta = ""
+			)
+
+		kp2.metric(
+			label= ":red[**⏳ Average age**]",
+			value = f"{average_age:.2f}" ,
+			delta = ""
+			)
+
+		kp3.metric(
+			label=":red[**✍ GCA-trimester 1**]",
+			value=f"{gca1:.2f}/20",
+			delta = 0.0
+			)
+		kp4.metric(
+			label=":red[**✍ GCA-trimester 2**]",
+			value = f"{gca2:.2f}/20",
+			delta = round(gca2-gca1, 3) 
+			)
+		kp5.metric(
+			label =":red[**✍ GCA-trimester 3**]",
+			value = f"{gca3:.2f}/20",
+			delta = round(gca3-gca2, 3) 
+			)
+		kp6.metric(
+			label =":red[**🎓 Annual-GCA**]",
+			value = f"{annual_gca:.2f}/20",
+			delta = "" 
+			)
+
+
+	placeholder2 = st.empty()
+	with placeholder2.container():
+		col1, col2 = st.columns(2)
+		col1.plotly_chart(pie_gender)
+		col1.caption('**We represent the gender of students in form 6e-5e (francophone section one of the cameroonian education system).**')
+		col2.plotly_chart(pie_form)
+		col2.caption("""**We represent the number of students in each form 6e and 5e**. 
+		**NB**: **Cameroon have two sub-education systems one is a francophone and a second is an anglosaxone.**   
+			""")
+	
+	tab1, tab2, tab3 = st.tabs([':orange[**Performance**]', ':orange[**Distribution**]', ':orange[**Miscelaneous**]'])
+
+	with tab1:
+		tab1.plotly_chart(bar_polar)
+		tab1.caption("""We represent the general class average of students in each sequence. Colors bar shows the marks 
+			for female gender and bar polar also shows the marks for male gender.""")
+		tab1.subheader('Progression')
+		tab1.line_chart(progression)
+		tab1.caption("""This chart shows a progression of students during the six evaluations. 
+			We just make a difference between the previous sequence and the next sequence. 
+			The x-axis represent a sequence and the y-axis represent the growth of students.""")
+
+		tab1.subheader('Passed or Failed')
+		col1, col2 = tab1.columns(2)
+		col1.caption(':red[💃 passed or failed table 👉]')
+		col1.dataframe(pass_or_fail)
+		col1.caption("""
+		- In left: We got this table to compute the number of students that the mark is  >10 (passed) or <10 (failed) for 
+		each sequence.
+		- In right: We plot this table. The chart explains how a student make some effort to succeed a ICT course.   
+		- In general, we can appreciate the effort of the students in the form 6e and 5e for each evaluation.   
+			""")
+		col2.bar_chart(pass_or_fail.T)
+
+		tab1.subheader("Student grades")
+		col3, col4 = tab1.columns(2)
+		col3.caption(':red[💃 grades table 👉]')
+		col3.dataframe(student_grade.fillna(0))
+		col3.caption("""   
+			The student grade respect this decision:
+			- U -> [0 - 5[;   E -> [5 - 9[
+			- D -> [9 - 12[;  C -> [12 - 15[
+			- B -> [15 - 18[; A -> [18 - 20[
+			""")
+		col4.bar_chart(student_grade.T)
+		col4.caption('This bar chart shows the number of student in each grade for each sequence.')
+
+	with tab2:
+		tab2.subheader('evaluation and age distribution')
+		tab2.caption('Histogram')
+		var1 = tab2.selectbox('Choose items', evaluations+['age'], key=6)
+		fig1 = px.histogram(data_frame=data, x=var1,  width=1200, height=500, opacity=0.75)
+		tab2.plotly_chart(fig1)
+		col1, col2 = tab2.columns(2)
+		col1.caption('Boxplot'); col2.caption('Violin')
+		var2 = col1.selectbox('Choose items', evaluations+['age'], key=7)
+		var3 = col2.selectbox('Choose items', evaluations+['age'], key=8)
+		col1.plotly_chart(px.box(data_frame=data, y=var2,  width=500, height=500))
+		col2.plotly_chart(px.violin(data_frame=data, x=var3,  width=500, height=500))
+
+
+	with tab3:
+		st.subheader('Relation graph')
+		container = tab3.empty()
+		vcol1, vcol2, vcol3 = container.columns(3)
+
+		trim1 = px.scatter(data_frame=df_trim, x="trimester 1", y="trimester 2", width=350, 
+			height=350,title="trimester 1 & 2")
+		vcol1.plotly_chart(trim1)
+
+		trim2 = px.scatter(data_frame=df_trim, x="trimester 2", y="trimester 3", width=350, height=350,title="trimester 3 & 2")
+		vcol2.plotly_chart(trim2)
+
+		trim3 = px.scatter(data_frame=df_trim, x="trimester 3", y="trimester 1", width=350, height=350,title="trimester 1 & 3")
+		vcol3.plotly_chart(trim3)
+
+		tab3.caption("""
+			The three charts shows the monotony of the relation function between the three trimesters. 
+			Each chart prove that the students for form 6e-5e francophone education systems at the Tebap college make considerably
+			an effort to succeed an ICT's course.
+			""")
+		tab3.subheader('Supervised segmentation with tree')
+		tab3.caption(""" 
+			In this section, we are making a supervised segmentation to segment the population of student into subgroups
+			that have different values for the target gender. To find a subgroups, we are using a tree structured model.
+			We cannot make a classification here because our dataset is just a size equal to 46. Let's go! 💂
+			""")
+		block = tab3.empty()
+		hcol1, hcol2 = block.columns(2)
+		hcol1.caption('**Feature importance table**👉')
+		hcol1.dataframe(feature_importances)
+		hcol1.caption("""  
+			The table shows the features that the tree structured model 
+			consider very importance for segmenting students
+			population in to subgroups. 
+			""")
+
+		hcol2.plotly_chart(bar_feature_importance)
+		tab3.subheader('Dot graph')
+		tab3.graphviz_chart(dot_graph)
+		tab3.caption("""
+			Let's us interprete this graph.
+
+			- node 0: the student that the mark in seq 6 <= 11.375 are female gender; the answer is False. In each node, we have 
+			a condition where the next node give an answers.
+
+			Let's plot the line separating the region.  
+			""")
+
+		bcol1, bcol2 = tab3.columns(2)
+
+		
+		var1 = bcol1.selectbox("Choose x-axis (line[i])", ['seq 1', "seq 2", "seq 4", "seq 6"])
+		var2 = bcol1.selectbox("Choose y-axis (line[i+1])", reversed(['seq 1', "seq 2", "seq 4", "seq 6"]))
+		
+		line1 = bcol2.multiselect("Line separator for x-axis", [1.25, 2.75, 3.125, 4.375,5.0, 5.625, 11.375,13.5]) 
+		line2 = bcol2.multiselect("Line separator for y-axis", [1.25, 2.75, 3.125, 4.375,5.0, 5.625, 11.375,13.5])
+		tab3.caption("""  
+			We have eight lines line 1 (node 0) to line 8 (node 12). We start to line 1 and line 2.  
+			""")
+		gplot = scatterplot(female, male, line_separator=[line1, line2], x_axis=var1, y_axis=var2)
+		tab3.pyplot(gplot)
+
+		
+
+		
+

module/dataset.py

@@ -0,0 +1,243 @@
+import streamlit as st 
+import numpy as np
+import pandas as pd
+from module.load_data import data
+from scipy import stats
+
+
+
+def create_data(data:pd.DataFrame) -> pd.DataFrame:
+	df = pd.DataFrame({"trimester 1": round((data['seq 1']+data['seq 2'])/2, 2),  
+		"trimester 2": round((data['seq 3']+data['seq 4'])/2, 2),
+		"trimester 3": round((data['seq 5']+data['seq 6'])/2, 2),
+		"age":data["age"]})
+	return df
+
+def subdata(data:pd.DataFrame, items:list) -> pd.DataFrame:
+	if len(items) == 3:
+		return data[(data[items[0][0]] == items[0][1]) & (data[items[1][0]] == items[1][1]) & (data[items[2][0]] == items[2][1])]['average_marks']
+
+	if len(items) == 2:
+		return data[(data[items[0][0]] == items[0][1]) & (data[items[1][0]] == items[1][1])]['average_marks']
+
+	if len(items) == 1:
+			return data[(data[items[0][0]] == items[0][1])]['average_marks']
+			
+
+def hypothesis_testing(group1:pd.DataFrame, group2:pd.DataFrame, label1:list, label2:list) -> list:
+
+	mean1, std1, nobs1 = group1.mean(), group1.std(), group1.count()
+	mean2, std2, nobs2 = group2.mean(), group2.std(), group2.count()
+
+	res = stats.ttest_ind_from_stats(mean1=mean1, std1=std1, nobs1=nobs1, 
+		mean2=mean2, std2=std2, nobs2=nobs2,equal_var=False)
+
+	tvalue = res.statistic
+	pvalue = res.pvalue 
+
+	diff = mean1 - mean2
+	std_error = np.sqrt((std1**2/nobs1) + (std2**2/nobs2))
+	low_bound   = diff - tvalue*std_error
+	upper_bound = diff + tvalue*std_error
+
+	text1 = f"""
+	:green[**Welch two-sample t-test**]
+
+	**data**: group1={label1} and group2={label2}
+
+	t = {tvalue:.4f}, diff = {diff:.4f}, p-value = {pvalue:.4f}
+
+	**alternative hypothesis**: true difference in means is not equal to 0
+
+	**95% confidence interval**: ({low_bound}, {upper_bound})
+
+	**mean of group1**: {mean1:.3f}
+
+	**mean of group2**: {mean2:.3f}
+	"""
+
+	text2 = f"""
+	:green[**Welch two-sample t-test**]
+
+	**data**: group1={label1} and group2={label2}
+
+	t = {tvalue:.8f}, diff = {diff:.8f}, p-value = {pvalue:.8f}
+
+	**null hypothesis**: true difference in means is equal to 0
+
+	**mean of group1**: {mean1:.3f}
+
+	**mean of group2**: {mean2:.3f}
+	"""
+
+	return text1 if pvalue < 0.05 else text2
+
+
+
+def eda():
+	df = pd.DataFrame({"trimester 1": round((data['seq 1']+data['seq 2'])/2, 2),  
+		"trimester 2": round((data['seq 3']+data['seq 4'])/2, 2),
+		"trimester 3": round((data['seq 5']+data['seq 6'])/2, 2),
+		"age":data["age"]})
+	corr = df.corr()
+
+	vif  = np.linalg.inv(corr.to_numpy()).diagonal()
+	vifs = pd.Series(np.round(vif,2), index=df.columns.tolist(), name="VIF") 
+
+	female = data[data.gender == 'F']
+	male   = data[data.gender == 'M']
+
+	df_female = create_data(female)
+	df_male = create_data(male)
+
+	items = ['F','M','6e','5e']
+
+
+	st.subheader('Sequence')
+	placeholder1 = st.empty()
+	
+	with placeholder1.container():
+		col1, col2 = st.columns(2)
+		col1.caption(':red[**Descriptive statistics**]')
+		col1.dataframe(round(data.describe(), 2))
+		col2.caption(':red[**Correlation**]')
+		col2.dataframe(round(data.corr(), 2))
+
+	st.subheader("Trimester")
+	st.caption(':red[**Descriptive statistics**]')
+	st.dataframe(round(df.describe(), 2), use_container_width=True)
+	placeholder2 = st.empty()
+	with placeholder2.container():
+		col1, col2 = st.columns(2)
+		col1.caption(':red[**Correlation**]')
+		col1.dataframe(round(corr, 2))
+		col2.caption(':red[**Variance Inflation Factor**]')
+		col2.dataframe(vifs)
+
+	with st.expander("👁 Read more"):
+		st.markdown(""" 
+			> 1. :orange[Descriptive statistics] help to define mean, standard deviation, minimun, maximun, median, etc.. 
+			such that we can summarize the dataset and discover the patterns.
+			> 2. :orange[Correlation] help to find the tendance or colinearity between two or more attributes in the dataset. 
+			> 3. :orange[VIF] is a mesure of colinearity among predictor variables within a multiple regression. 
+			>> 1. If outcome is 1, it's okay.
+			>> 2. If it is between 1 and 5, it show low to average colinearity, and above 5 generally means highly redundant
+			and variable should be dropped. 
+			""")
+
+	st.subheader('Sequence by gender')
+	placeholder3 = st.empty()
+	placeholder4 = st.empty()
+	with placeholder3.container():
+		col1, col2 = st.columns(2)
+		col1.caption(':orange[**Descriptive statistics: Female**]')
+		col1.dataframe(round(female.describe(), 2))
+		col2.caption(':orange[**Descriptive statistics: Male**]')
+		col2.dataframe(round(male.describe(), 2))
+
+	with placeholder4.container():
+		col1, col2 = st.columns(2)
+		col1.caption(':orange[**Correlation: Female**]')
+		col1.dataframe(round(female.corr(), 2))
+		col2.caption(':orange[**Correlation: Male**]')
+		col2.dataframe(round(male.corr(), 2))
+
+	st.subheader('Trimester by gender')
+	placeholder5 = st.empty()
+	placeholder6 = st.empty()
+	with placeholder5.container():
+		col1, col2 = st.columns(2)
+		col1.caption(':orange[**Descriptive statistics: Female**]')
+		col1.dataframe(round(df_female.describe(), 2))
+		col2.caption(':orange[**Descriptive statistics: Male**]')
+		col2.dataframe(round(df_male.describe(), 2))
+
+	with placeholder6.container():
+		col1, col2 = st.columns(2)
+		col1.caption(':orange[**Correlation: Female**]')
+		col1.dataframe(round(df_female.corr(), 2))
+		col2.caption(':orange[**Correlation: Male**]')
+		col2.dataframe(round(df_male.corr(), 2))
+
+	st.subheader('Assumption')
+	st.markdown("""
+		In this section, you can make your assumption to know which group is best than other group in this ICT course.
+		For example: According to table trimester evaluation by gender, we have two groups (female students and male students). 
+
+		By the observation, 
+		> **Is it true that female students are best than male students in the ICT course for Tebap college?**  
+
+		This question we allow us to compare the general annual class average (mean) for one group to other group. That's lead to 
+		compute the difference between the two means. We have two hypothesis:
+
+		1. **Null hypothesis:** :blue[ true difference in means is equal to 0].
+
+		2. **Alternative hypothesis:** :blue[ true difference in means is not equal to 0]
+
+		**NB**: difference = $\mu$(group1) - $\mu$(group2)
+
+		We choose the right hypothesis as follows:
+		- **if p-value < 5% then we reject null hypothesis**
+		- **if p-value > 5% then we accept null hypothesis**
+		""")
+	placeholder7 = st.empty()
+	with placeholder7.container():
+		col1, col2 = st.columns(2)
+		
+		with col1:
+			label1 = []
+			st.caption(':red[Create group A]')
+			scol1, scol2, scol3 = st.columns(3)
+			
+			gender1 = scol1.selectbox('Gender', ['Choose an option', 'F', 'M'])
+			form1    = scol2.selectbox('Form',  ['Choose an option', '6e', '5e'])
+			age1 = scol3.selectbox('Age', ['Choose an option',]+sorted(data.age.unique().tolist()))
+
+			if gender1 != 'Choose an option':
+				label1.append(('gender', gender1))
+			if form1 != 'Choose an option':
+				label1.append(('form', form1))
+			if age1 != 'Choose an option':
+				label1.append(('age', age1))
+			
+			group1 = subdata(data, label1)
+			if type(group1) == pd.Series:
+				scol2.write(group1)
+			
+
+		with col2:
+			label2 = []
+			st.caption(':red[Create group B]')
+			tcol1, tcol2, tcol3 = st.columns(3)
+			
+			gender2 = tcol1.selectbox('Gender', ['Choose an option', 'F', 'M'], key=1)
+			form2    = tcol2.selectbox('Form',  ['Choose an option', '6e', '5e'], key=2)
+			age2 = tcol3.selectbox('Age', ['Choose an option',]+sorted(data.age.unique().tolist()), key=3)
+
+			if gender2 != 'Choose an option':
+				label2.append(('gender', gender2))
+			if form2 != 'Choose an option':
+				label2.append(('form', form2))
+			if age2 != 'Choose an option':
+				label2.append(('age', age2))
+			
+			group2 = subdata(data, label2)
+			if type(group2) == pd.Series:
+				tcol2.write(group2)
+
+		st.caption('Choose only one or two item(s) per group. For example group A = (F, 5e) or group B = (M)')
+
+		if (sorted(label2) == sorted(label1)) and not(len(label1) == 0 or len(label2) == 0):
+			st.info(f'You cannot create two different groups with same items: label1 = {label1} & label2 = {label2}.')
+
+
+		if st.button('t-test'):
+			res = hypothesis_testing(group1, group2, label1, label2)
+			st.write(res)
+
+		
+
+
+
+
+

module/load_data.py

@@ -0,0 +1,9 @@
+import pandas as pd
+
+
+def get_data():
+	data = pd.read_excel('data/Note_des_classes_6e_5e.xlsx')
+	data['average_marks'] = round(data['average_marks'], 2)
+	return data
+
+data = get_data()

requirements.txt

@@ -0,0 +1,9 @@
+scipy==1.7.3
+matplotlib==3.5.3
+plotly==5.14.0
+numpy==1.21.6
+pandas==1.3.5
+streamlit==1.23.1
+Pillow
+openpyxl
+scikit_learn==1.3.2