andwati/predictive_analysis
1
0
Fork 0
mirror of https://github.com/andwati/predictive_analysis.git synced 2026-11-02 23:22:15 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add Streamlit app for student success prediction and model training script

Browse Source 
- Created `streamlit_app.py` for a user-friendly dashboard to predict student outcomes using a Random Forest model.
- Added `requirements.txt` to specify necessary Python packages for the application.
- Implemented `train_and_save_model.py` to train the Random Forest model and save it along with preprocessing information.
- Included detailed user input forms for various student characteristics and academic performance metrics.
- Added custom styling for better UI/UX in the Streamlit app.
- Integrated prediction results display with probability metrics and model performance insig
This commit is contained in:
andwati
2025-09-27 23:45:14 +03:00
commit ed2ff5b359
7 changed files with 19386 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
streamlit_app.py
+547
View File
@@ -0,0 +1,547 @@
"""
Student Success Prediction Dashboard
A beautiful Streamlit app for predicting student academic outcomes using the trained Random Forest model.
"""
import streamlit as st
import pandas as pd
import numpy as np
import pickle
import plotly.express as px
import plotly.graph_objects as go
from plotly.subplots import make_subplots
# Configure Streamlit page
st.set_page_config(
page_title="🎓 Student Success Predictor",
page_icon="🎓",
layout="wide",
initial_sidebar_state="expanded"
)
# Custom CSS for beautiful styling
st.markdown("""
<style>
.main-header {
font-size: 3rem;
color: #1f77b4;
text-align: center;
margin-bottom: 2rem;
font-weight: bold;
}
.sub-header {
font-size: 1.5rem;
color: #2c3e50;
margin-bottom: 1rem;
}
.prediction-box {
padding: 2rem;
border-radius: 10px;
text-align: center;
font-size: 1.5rem;
font-weight: bold;
margin: 1rem 0;
}
.graduate-box {
background: linear-gradient(135deg, #28a745, #20c997);
color: white;
}
.dropout-box {
background: linear-gradient(135deg, #dc3545, #fd7e14);
color: white;
}
.enrolled-box {
background: linear-gradient(135deg, #ffc107, #fd7e14);
color: white;
}
.metric-card {
background: #f8f9fa;
padding: 1rem;
border-radius: 8px;
border-left: 4px solid #1f77b4;
}
</style>
""", unsafe_allow_html=True)
@st.cache_data
def load_model_and_info():
"""Load the trained model and preprocessing information"""
try:
with open('random_forest_model.pkl', 'rb') as f:
model = pickle.load(f)
with open('preprocessing_info.pkl', 'rb') as f:
preprocessing_info = pickle.load(f)
with open('feature_names.pkl', 'rb') as f:
feature_names = pickle.load(f)
return model, preprocessing_info, feature_names
except FileNotFoundError:
st.error("❌ Model files not found! Please run 'train_and_save_model.py' first.")
st.stop()
def create_feature_mappings():
"""Create human-readable mappings for categorical features"""
# Marital status mapping
marital_status_map = {
1: "Single",
2: "Married",
3: "Widowed",
4: "Divorced",
5: "Civil Union",
6: "Legally Separated"
}
# Application mode mapping (simplified for demo)
application_mode_map = {
1: "1st Phase - General Contingent",
2: "Ordinance No. 612/93",
5: "1st Phase - Special Contingent (Azores)",
7: "Holders of Other Higher Courses",
10: "Ordinance No. 854-B/99",
15: "International Student (Bachelor)",
16: "1st Phase - Special Contingent (Madeira)",
17: "2nd Phase - General Contingent",
18: "3rd Phase - General Contingent",
26: "Ordinance No. 533-A/99, item b2",
27: "Ordinance No. 533-A/99, item b3",
39: "Over 23 years old",
42: "Transfer",
43: "Change of course",
44: "Technological specialization diploma holders",
51: "Change of institution/course",
53: "Short cycle diploma holders",
57: "Change of institution/course (International)"
}
# Course mapping (simplified - showing most common ones)
course_map = {
33: "Biofuel Production Technologies",
171: "Animation and Multimedia Design",
8014: "Social Service (evening attendance)",
9003: "Agronomy",
9070: "Communication Design",
9085: "Veterinary Nursing",
9119: "Informatics Engineering",
9130: "Equinculture",
9147: "Management",
9238: "Social Service",
9254: "Tourism",
9500: "Nursing",
9556: "Oral Hygiene",
9670: "Advertising and Marketing Management",
9773: "Journalism and Communication",
9853: "Basic Education",
9991: "Management (evening attendance)"
}
# Previous qualification mapping
qualification_map = {
1: "Secondary education",
2: "Higher education - bachelor's degree",
3: "Higher education - degree",
4: "Higher education - master's",
5: "Higher education - doctorate",
6: "Frequency of higher education",
9: "12th year of schooling - not completed",
10: "11th year of schooling - not completed",
12: "Other - 11th year of schooling",
14: "10th year of schooling",
15: "10th year of schooling - not completed",
19: "Basic education 3rd cycle (9th/10th/11th year) or equiv.",
38: "Basic education 2nd cycle (6th/7th/8th year) or equiv.",
39: "Technological specialization course",
40: "Higher education - degree (1st cycle)",
42: "Professional course",
43: "Higher education - master (2nd cycle)"
}
# Occupation mapping (simplified)
occupation_map = {
0: "Student",
1: "Representatives of the Legislative Power and Executive Bodies",
2: "Specialists in Intellectual and Scientific Activities",
3: "Intermediate Level Technicians and Professions",
4: "Administrative staff",
5: "Personal Services, Security and Safety Workers",
6: "Farmers and Skilled Workers in Agriculture",
7: "Skilled Workers in Industry, Construction and Craftsmen",
8: "Installation and Machine Operators and Assembly Workers",
9: "Unskilled Workers",
10: "Armed Forces Professions",
90: "Other Situation",
99: "Unspecified",
122: "Health professionals",
123: "Teachers",
125: "Specialists in information and communication technologies",
131: "Intermediate level science and engineering technicians",
132: "Technicians and associate professionals, health",
134: "Intermediate level technicians from legal, social, sports",
141: "Office workers, secretaries in general and data processing",
143: "Data, accounting, statistical, financial services",
144: "Other administrative support staff",
151: "Personal service workers",
152: "Sellers",
153: "Personal care workers and the like",
171: "Skilled construction workers and the like, except electricians",
173: "Skilled workers in printing, precision instrument manufacturing",
175: "Workers in food processing, woodworking, clothing",
191: "Cleaning workers",
192: "Unskilled workers in agriculture, animal production",
193: "Unskilled workers in extractive industry, construction",
194: "Meal preparation assistants"
}
return {
'marital_status': marital_status_map,
'application_mode': application_mode_map,
'course': course_map,
'qualification': qualification_map,
'occupation': occupation_map
}
def preprocess_input(user_input, preprocessing_info):
"""Preprocess user input to match the training data format"""
# Create a DataFrame with the user input
input_df = pd.DataFrame([user_input])
# Get categorical columns that need one-hot encoding
categorical_columns = [
'Marital status', 'Application mode', 'Course', 'Daytime/evening attendance',
'Previous qualification', 'Nationality', 'Mothers qualification',
'Fathers qualification', 'Mothers occupation', 'Fathers occupation',
'Gender', 'Displaced', 'Educational special needs', 'Debtor',
'Tuition fees up to date', 'Scholarship holder', 'International',
'Application order'
]
# Perform one-hot encoding for categorical columns
for column in categorical_columns:
if column in input_df.columns:
# Get dummies with prefix
dummies = pd.get_dummies(input_df[column], prefix=column)
# Concatenate with original data
input_df = pd.concat([input_df, dummies], axis=1)
# Drop original column
input_df.drop(columns=[column], inplace=True)
# Ensure all training features are present
feature_names = preprocessing_info['feature_names']
# Add missing columns with 0 values
for feature in feature_names:
if feature not in input_df.columns:
input_df[feature] = 0
# Select only the features used in training and in the same order
input_df = input_df[feature_names]
return input_df
def main():
"""Main Streamlit application"""
# Load model and preprocessing info
model, preprocessing_info, feature_names = load_model_and_info()
# Create feature mappings for user-friendly display
feature_maps = create_feature_mappings()
# Header
st.markdown('<h1 class="main-header">🎓 Student Success Predictor</h1>', unsafe_allow_html=True)
st.markdown("**Predict whether a student will Graduate, Dropout, or remain Enrolled based on academic and demographic factors.**")
# Sidebar for input
st.sidebar.markdown('<h2 class="sub-header">📝 Student Information</h2>', unsafe_allow_html=True)
# Create input form
with st.sidebar:
# Personal Information
st.markdown("### 👤 Personal Information")
marital_status = st.selectbox(
"Marital Status",
options=list(feature_maps['marital_status'].keys()),
format_func=lambda x: feature_maps['marital_status'][x]
)
gender = st.selectbox("Gender", [0, 1], format_func=lambda x: "Female" if x == 0 else "Male")
age_at_enrollment = st.slider("Age at Enrollment", 17, 70, 20)
nationality = st.selectbox("Nationality", [1, 2, 6, 11, 13, 14, 17, 21, 22, 24, 25, 26, 32, 41, 62, 100, 101, 103, 105, 108, 109])
# Academic Background
st.markdown("### 🎓 Academic Background")
application_mode = st.selectbox(
"Application Mode",
options=list(feature_maps['application_mode'].keys()),
format_func=lambda x: feature_maps['application_mode'].get(x, f"Mode {x}")
)
application_order = st.slider("Application Order", 0, 9, 1)
course = st.selectbox(
"Course",
options=list(feature_maps['course'].keys()),
format_func=lambda x: feature_maps['course'].get(x, f"Course {x}")
)
daytime_evening = st.selectbox("Attendance", [0, 1], format_func=lambda x: "Evening" if x == 0 else "Daytime")
previous_qualification = st.selectbox(
"Previous Qualification",
options=list(feature_maps['qualification'].keys()),
format_func=lambda x: feature_maps['qualification'].get(x, f"Qualification {x}")
)
previous_qualification_grade = st.slider("Previous Qualification Grade", 95, 190, 120)
admission_grade = st.slider("Admission Grade", 95, 190, 120)
# Family Background
st.markdown("### 👨‍👩‍👧‍👦 Family Background")
mothers_qualification = st.selectbox(
"Mother's Qualification",
options=list(feature_maps['qualification'].keys()),
format_func=lambda x: feature_maps['qualification'].get(x, f"Qualification {x}")
)
fathers_qualification = st.selectbox(
"Father's Qualification",
options=list(feature_maps['qualification'].keys()),
format_func=lambda x: feature_maps['qualification'].get(x, f"Qualification {x}")
)
mothers_occupation = st.selectbox(
"Mother's Occupation",
options=list(feature_maps['occupation'].keys()),
format_func=lambda x: feature_maps['occupation'].get(x, f"Occupation {x}")
)
fathers_occupation = st.selectbox(
"Father's Occupation",
options=list(feature_maps['occupation'].keys()),
format_func=lambda x: feature_maps['occupation'].get(x, f"Occupation {x}")
)
# Financial and Social Factors
st.markdown("### 💰 Financial & Social")
displaced = st.selectbox("Displaced", [0, 1], format_func=lambda x: "No" if x == 0 else "Yes")
educational_special_needs = st.selectbox("Educational Special Needs", [0, 1], format_func=lambda x: "No" if x == 0 else "Yes")
debtor = st.selectbox("Debtor", [0, 1], format_func=lambda x: "No" if x == 0 else "Yes")
tuition_fees_up_to_date = st.selectbox("Tuition Fees Up to Date", [0, 1], format_func=lambda x: "No" if x == 0 else "Yes")
scholarship_holder = st.selectbox("Scholarship Holder", [0, 1], format_func=lambda x: "No" if x == 0 else "Yes")
international = st.selectbox("International Student", [0, 1], format_func=lambda x: "No" if x == 0 else "Yes")
# Academic Performance - 1st Semester
st.markdown("### 📚 1st Semester Performance")
curricular_units_1st_sem_credited = st.slider("Curricular Units 1st Sem (Credited)", 0, 26, 0)
curricular_units_1st_sem_enrolled = st.slider("Curricular Units 1st Sem (Enrolled)", 0, 26, 6)
curricular_units_1st_sem_evaluations = st.slider("Curricular Units 1st Sem (Evaluations)", 0, 45, 6)
curricular_units_1st_sem_approved = st.slider("Curricular Units 1st Sem (Approved)", 0, 26, 6)
curricular_units_1st_sem_grade = st.slider("Curricular Units 1st Sem (Grade)", 0, 20, 13)
curricular_units_1st_sem_without_evaluations = st.slider("Curricular Units 1st Sem (Without Evaluations)", 0, 14, 0)
# Academic Performance - 2nd Semester
st.markdown("### 📖 2nd Semester Performance")
curricular_units_2nd_sem_credited = st.slider("Curricular Units 2nd Sem (Credited)", 0, 20, 0)
curricular_units_2nd_sem_enrolled = st.slider("Curricular Units 2nd Sem (Enrolled)", 0, 23, 6)
curricular_units_2nd_sem_evaluations = st.slider("Curricular Units 2nd Sem (Evaluations)", 0, 33, 6)
curricular_units_2nd_sem_approved = st.slider("Curricular Units 2nd Sem (Approved)", 0, 20, 6)
curricular_units_2nd_sem_grade = st.slider("Curricular Units 2nd Sem (Grade)", 0, 20, 13)
curricular_units_2nd_sem_without_evaluations = st.slider("Curricular Units 2nd Sem (Without Evaluations)", 0, 12, 0)
# Economic Indicators
st.markdown("### 📊 Economic Indicators")
unemployment_rate = st.slider("Unemployment Rate (%)", 7.0, 16.2, 10.8)
inflation_rate = st.slider("Inflation Rate (%)", -0.8, 3.7, 1.4)
gdp = st.slider("GDP", -4.06, 3.51, 1.74)
# Create user input dictionary
user_input = {
'Marital status': marital_status,
'Application mode': application_mode,
'Application order': application_order,
'Course': course,
'Daytime/evening attendance': daytime_evening,
'Previous qualification': previous_qualification,
'Previous qualification (grade)': previous_qualification_grade,
'Nationality': nationality,
'Mothers qualification': mothers_qualification,
'Fathers qualification': fathers_qualification,
'Mothers occupation': mothers_occupation,
'Fathers occupation': fathers_occupation,
'Admission grade': admission_grade,
'Displaced': displaced,
'Educational special needs': educational_special_needs,
'Debtor': debtor,
'Tuition fees up to date': tuition_fees_up_to_date,
'Gender': gender,
'Scholarship holder': scholarship_holder,
'Age at enrollment': age_at_enrollment,
'International': international,
'Curricular units 1st sem (credited)': curricular_units_1st_sem_credited,
'Curricular units 1st sem (enrolled)': curricular_units_1st_sem_enrolled,
'Curricular units 1st sem (evaluations)': curricular_units_1st_sem_evaluations,
'Curricular units 1st sem (approved)': curricular_units_1st_sem_approved,
'Curricular units 1st sem (grade)': curricular_units_1st_sem_grade,
'Curricular units 1st sem (without evaluations)': curricular_units_1st_sem_without_evaluations,
'Curricular units 2nd sem (credited)': curricular_units_2nd_sem_credited,
'Curricular units 2nd sem (enrolled)': curricular_units_2nd_sem_enrolled,
'Curricular units 2nd sem (evaluations)': curricular_units_2nd_sem_evaluations,
'Curricular units 2nd sem (approved)': curricular_units_2nd_sem_approved,
'Curricular units 2nd sem (grade)': curricular_units_2nd_sem_grade,
'Curricular units 2nd sem (without evaluations)': curricular_units_2nd_sem_without_evaluations,
'Unemployment rate': unemployment_rate,
'Inflation rate': inflation_rate,
'GDP': gdp
}
# Main content area
col1, col2 = st.columns([2, 1])
with col1:
st.markdown('<h2 class="sub-header">🔮 Prediction Results</h2>', unsafe_allow_html=True)
# Make prediction
if st.button("🚀 Predict Student Outcome", type="primary"):
with st.spinner("Analyzing student data..."):
# Preprocess input
processed_input = preprocess_input(user_input, preprocessing_info)
# Make prediction
prediction = model.predict(processed_input)[0]
prediction_proba = model.predict_proba(processed_input)[0]
# Get prediction label
target_reverse_mapping = preprocessing_info['target_reverse_mapping']
predicted_outcome = target_reverse_mapping[prediction]
# Display prediction with styling
if predicted_outcome == "Graduate":
st.markdown(f'<div class="prediction-box graduate-box">🎓 Predicted Outcome: GRADUATE</div>', unsafe_allow_html=True)
elif predicted_outcome == "Dropout":
st.markdown(f'<div class="prediction-box dropout-box">⚠️ Predicted Outcome: DROPOUT</div>', unsafe_allow_html=True)
else:
st.markdown(f'<div class="prediction-box enrolled-box">📚 Predicted Outcome: ENROLLED</div>', unsafe_allow_html=True)
# Display probabilities
st.markdown("### 📊 Prediction Confidence")
prob_df = pd.DataFrame({
'Outcome': ['Dropout', 'Enrolled', 'Graduate'],
'Probability': prediction_proba
})
# Create probability chart
fig = px.bar(
prob_df,
x='Outcome',
y='Probability',
color='Outcome',
color_discrete_map={
'Graduate': '#28a745',
'Dropout': '#dc3545',
'Enrolled': '#ffc107'
},
title="Prediction Probabilities"
)
fig.update_layout(showlegend=False)
st.plotly_chart(fig, use_container_width=True)
# Display detailed probabilities
col_a, col_b, col_c = st.columns(3)
with col_a:
st.markdown(f"""
<div class="metric-card">
<h4>🎓 Graduate</h4>
<h2>{prediction_proba[2]:.1%}</h2>
</div>
""", unsafe_allow_html=True)
with col_b:
st.markdown(f"""
<div class="metric-card">
<h4>⚠️ Dropout</h4>
<h2>{prediction_proba[0]:.1%}</h2>
</div>
""", unsafe_allow_html=True)
with col_c:
st.markdown(f"""
<div class="metric-card">
<h4>📚 Enrolled</h4>
<h2>{prediction_proba[1]:.1%}</h2>
</div>
""", unsafe_allow_html=True)
with col2:
st.markdown('<h2 class="sub-header">ℹ️ About the Model</h2>', unsafe_allow_html=True)
st.info("""
**🤖 Model Details:**
- **Algorithm:** Random Forest Classifier
- **Features:** 36 student characteristics
- **Training Accuracy:** 100%
- **Test Accuracy:** 93.84%
- **Classes:** Graduate, Dropout, Enrolled
""")
st.markdown("### 📈 Key Factors")
st.markdown("""
The model considers:
- **Academic Performance** (grades, units)
- **Demographics** (age, family background)
- **Economic Indicators** (GDP, unemployment)
- **Social Factors** (scholarships, special needs)
""")
st.markdown("### 🎯 Use Cases")
st.markdown("""
- **Early Warning System** for at-risk students
- **Resource Allocation** planning
- **Academic Support** targeting
- **Institutional Planning** insights
""")
# Model performance visualization
st.markdown("### 📊 Model Performance")
# Create a simple performance chart
metrics = ['Accuracy', 'Precision', 'Recall', 'F1-Score']
values = [93.84, 92.96, 94.02, 93.15] # From the notebook results
fig_perf = go.Figure(data=go.Scatterpolar(
r=values,
theta=metrics,
fill='toself',
line_color='#1f77b4'
))
fig_perf.update_layout(
polar=dict(
radialaxis=dict(
visible=True,
range=[0, 100]
)),
showlegend=False,
title="Model Performance Metrics",
height=300
)
st.plotly_chart(fig_perf, use_container_width=True)
# Footer
st.markdown("---")
st.markdown("*Developed with ❤️ using Streamlit and scikit-learn*")
if __name__ == "__main__":
main()