Upload 37 files

app.py

@@ -4,6 +4,7 @@ import streamlit as st
 import eda
 import prediction
 
+# Set the page configuration for the Streamlit app
 st.set_page_config(
     page_title="Waste Classification",
     page_icon="♻️",
@@ -12,13 +13,15 @@ st.set_page_config(
 )
 
 def main():
+    # Create a sidebar for navigation
     st.sidebar.title("♻️ Navigation")
     page = st.sidebar.radio("Go to", ["🏠 Home", "📊 EDA", "🔍 Prediction"])
 
     if page == "🏠 Home":
+        # Add sidebar content for the Home page
         st.sidebar.markdown("---")
         st.sidebar.subheader("📊 About the Model")
-        accuracy = 0.80
+        accuracy = 0.82
         st.sidebar.write("🎯 Model Accuracy:")
         st.sidebar.progress(accuracy)
         st.sidebar.write(f"{accuracy:.2%}")
@@ -28,9 +31,23 @@ def main():
         st.sidebar.write(f"Our model correctly classifies {accuracy:.2%} of waste items, helping improve recycling efficiency.")
 
         st.sidebar.markdown("---")
-        st.sidebar.subheader("♻️ Fun Fact")
-        st.sidebar.info("Proper waste classification can increase recycling rates by up to 50%!")
+        st.sidebar.subheader("♻️ Fun Facts")
+        # Added more fun facts
+        fun_facts = [
+            "Proper waste classification can increase recycling rates by up to 50%!",
+            "Recycling one aluminum can saves enough energy to run a TV for three hours.",
+            "It takes 450 years for a plastic bottle to decompose in a landfill.",
+            "Glass can be recycled endlessly without losing quality or purity.",
+            "Recycling paper saves 17 trees and 7,000 gallons of water per ton of paper."
+        ]
+        st.sidebar.info(fun_facts[st.session_state.get('fun_fact_index', 0)])
+        
+        # Button to cycle through fun facts
+        if st.sidebar.button("Next Fun Fact"):
+            st.session_state['fun_fact_index'] = (st.session_state.get('fun_fact_index', 0) + 1) % len(fun_facts)
+            st.experimental_rerun()
 
+        # Main content for the Home page
         st.title("♻️ Welcome to Waste Classification Tool")
         st.write("""
         This application provides functionalities for Exploratory Data Analysis and 
@@ -38,36 +55,61 @@ def main():
         select the module you wish to utilize.
         """)
         
+        # Display an image in the center column
         col1, col2, col3 = st.columns([1,2,1])
         with col2:
             st.image("https://assets-a1.kompasiana.com/items/album/2021/03/14/dr-stone-fandomcom-1536x864-604dff978ede483a3b589c96.png?t=o&v=780",  
-                     caption="Notes : Recycle is important for our environment", use_column_width=True)
+                     caption="Notes: Recycling is crucial for our environment", use_column_width=True)
         
         st.markdown("---")
         
+        # Information about the dataset
         st.write("#### 📊 Dataset")
         st.info("""
         The dataset used is the RealWaste dataset, containing images of waste items across 9 major material types, 
-        collected within an authentic landfill environment.
+        collected within an authentic landfill environment. This dataset provides a realistic representation of 
+        waste items, allowing our model to learn from real-world examples.
         """)
         
+        # Problem statement
         st.write("#### ⚠️ Problem Statement")
         st.warning("""
         Manual waste sorting is inefficient, leading to low recycling rates and increased environmental harm. 
         Our goal is to develop a deep learning-based waste classification system using a Convolutional Neural Network (CNN) 
-        that can accurately classify at least 70% of waste images across 9 material categories.
+        that can accurately classify at least 70% of waste images across 9 material categories. This automated 
+        system aims to significantly improve recycling efficiency and reduce the environmental impact of improper 
+        waste disposal.
         """)
         
+        # Project objective
         st.write("#### 🎯 Objective")
         st.success("""
-        This project aims to develop a CNN model capable of accurately classifying waste images into distinct material types, 
-        improving waste management efficiency and environmental sustainability.
+        This project aims to develop a Convolutional Neural Network (CNN) model capable of accurately classifying waste images into nine distinct material types:
+        1. Cardboard - e.g., boxes, packaging
+        2. Glass - e.g., bottles, jars
+        3. Metal - e.g., cans, foil
+        4. Paper - e.g., newspapers, magazines
+        5. Plastic - e.g., bottles, containers
+        6. Miscellaneous Trash - e.g., non-recyclable items
+        7. Food Organics - e.g., fruit peels, vegetable scraps
+        8. Textile Trash - e.g., old clothes, fabrics
+        9. Vegetation - e.g., leaves, branches
+
+        By leveraging deep learning techniques, we seek to automate and improve waste management efficiency, ultimately contributing to environmental sustainability. Our model will analyze visual characteristics such as shape, color, and texture to categorize waste items, helping to streamline recycling processes and reduce the environmental impact of improper waste disposal.
+
+        The successful implementation of this project could lead to:
+        - Increased recycling rates
+        - Reduced contamination in recycling streams
+        - Lower operational costs for waste management facilities
+        - Enhanced public awareness about proper waste sorting
         """)
 
     elif page == "📊 EDA":
+        # Run the Exploratory Data Analysis module
         eda.run()
     
     elif page == "🔍 Prediction":
+        # Run the Prediction module
         prediction.run()
 
 if __name__ == "__main__":

eda.py

@@ -1,5 +1,6 @@
 # eda.py
 
+# Import necessary libraries
 import streamlit as st
 import pandas as pd
 import plotly.express as px
@@ -7,6 +8,7 @@ import matplotlib.pyplot as plt
 import seaborn as sns
 import numpy as np
 
+# Function to load data with caching for performance
 @st.cache_data
 def load_data():
     return pd.DataFrame({
@@ -14,33 +16,74 @@ def load_data():
         'Number of Images': [921, 790, 500, 495, 461, 436, 420, 411, 318]
     })
 
+# Main function to run the Streamlit app
 def run():
     st.title('📊 Exploratory Data Analysis - Waste Classification')
 
+    # Load the data
     data = load_data()
 
-    if st.checkbox('Show dataset info'):
-        st.subheader("Dataset Information")
-        st.write(data)
-        
-        st.write("The dataset shows an uneven distribution across the nine waste categories. "
-                 "This imbalance may impact model performance and will need to be addressed during the model training phase.")
-
-    st.subheader("Distribution of Waste Categories")
-    fig = px.bar(data, x='Class', y='Number of Images', color='Class',
-                 title='Number of Images per Waste Category')
-    st.plotly_chart(fig)
+    # Create a selectbox for users to choose visualization
+    visualization_option = st.selectbox(
+        "Choose a visualization:",
+        ("Dataset Information and Distribution", "Sample Images")
+    )
 
-    st.write("This chart shows the distribution of images across different waste categories. "
-             "Plastic and Metal categories have significantly more images, which could lead to bias in the model.")
+    if visualization_option == "Dataset Information and Distribution":
+        st.subheader("Dataset Information and Distribution")
+        
+        # Add checkbox for showing dataset information
+        show_dataset_info = st.checkbox("Show Dataset Information", value=True)
+        
+        if show_dataset_info:
+            st.write(data)
+            st.write("The dataset shows an uneven distribution across the nine waste categories. "
+                     "This imbalance may impact model performance and will need to be addressed during the model training phase.")
 
+        # Bar chart
+        fig_bar = px.bar(data, x='Class', y='Number of Images', color='Class',
+                     title='Number of Images per Waste Category')
+        st.plotly_chart(fig_bar, use_container_width=True)
 
-    st.subheader("Sample Images")
-    st.write("Here are sample images from each waste category:")
+        # Pie chart
+        fig_pie = px.pie(data, values='Number of Images', names='Class',
+                         title='Proportion of Images per Waste Category')
+        st.plotly_chart(fig_pie, use_container_width=True)
 
-    st.subheader("Image Size Distribution")
-    st.write("All images in the dataset have a resolution of 524x524 pixels.")
+        st.write("These charts show the distribution of images across different waste categories. "
+                 "Plastic and Metal categories have significantly more images, which could lead to bias in the model.")
 
+    elif visualization_option == "Sample Images":
+        st.subheader("Sample Images")
+        st.write("Here are sample images from each waste category:")
+        
+        categories = ['cardboard', 'food_organics', 'glass', 'metal', 'misc', 'paper', 'plastic', 'textile', 'vegetation']
+        
+        # Create a selectbox for choosing a specific category
+        selected_category = st.selectbox("Select a waste category:", categories)
+        
+        st.write(f"**{selected_category.capitalize()}**")
+        
+        cols = st.columns(3)
+        for i in range(1, 4):
+            with cols[i-1]:
+                img_path = f'./visualization/{selected_category} ({i}).jpg'
+                st.image(img_path, caption=f'{selected_category.capitalize()} ({i})', use_column_width=True)
+        
+        st.write("These sample images provide a visual representation of the selected waste category in our dataset.")
+        
+        # Add an option to view all categories
+        if st.checkbox("View all categories"):
+            for category in categories:
+                if category != selected_category:
+                    st.write(f"**{category.capitalize()}**")
+                    cols = st.columns(3)
+                    for i in range(1, 4):
+                        with cols[i-1]:
+                            img_path = f'./visualization/{category} ({i}).jpg'
+                            st.image(img_path, caption=f'{category.capitalize()} ({i})', use_column_width=True)
+                    st.markdown("---")  # Add a horizontal line after each category
 
+# Entry point of the script
 if __name__ == "__main__":
     run()