project.py

import streamlit as st
import pandas as pd
from sklearn.preprocessing import StandardScaler, LabelEncoder
from sklearn.manifold import TSNE
import matplotlib.pyplot as plt
import seaborn as sns

# ΒΗΜΑ 1ο: Φόρτωση Δεδομένων: Η εφαρμογή θα πρέπει να είναι σε θέση να φορτώνει tabular data (csv)
def load_data():
    uploaded_file = st.file_uploader("heart_failure_clinical_records.csv")
    if uploaded_file:
        if uploaded_file.name.endswith('.csv'):
            data = pd.read_csv(uploaded_file)
        else:
            data = pd.read_excel(uploaded_file)
        st.success("File loaded successfully!")    
        return data
      
# ΒΗΜΑ 2ο: Προδιαγραφές Πίνακα
def validate_data(data):
    if data is not None:
        rows, cols = data.shape
        if cols < 2:
            st.error("The data must have at least one feature column and one label column")
            return False
        
        # Assuming the last column is the label
        features = data.iloc[:, :-1]
        label = data.iloc[:, -1]
        
        if label.nunique() < 2:
            st.error("The label column must contain at least two unique values")
            return False
        
        st.success(f"Data has {rows} samples and {cols-1} features with 1 label column.")
        return True
    return False      

def preprocess_data(data):
    # Remove rows with any non-numeric data in features
    features = data.iloc[:, :-1]
    features = features.apply(pd.to_numeric, errors='coerce')
    label = data.iloc[:, -1]
    
    # Drop rows with NaN values
    data_clean = features.dropna()
    data_clean[label.name] = label.loc[data_clean.index]
    
    return data_clean

# ΒΗΜΑ 3ο: 2D Visualization Tab
def plot_tsne(data):
    if data.empty:
        st.error("No valid data available for t-SNE.")
        return
    
    features = data.iloc[:, :-1]
    labels = data.iloc[:, -1]
    tsne = TSNE(n_components=2, perplexity=30, n_iter=300)
    tsne_result = tsne.fit_transform(features)
    
    plt.figure(figsize=(10, 6))
    sns.scatterplot(x=tsne_result[:, 0], y=tsne_result[:, 1], hue=labels, palette='viridis')
    plt.title('t-SNE 2D Visualization')
    st.pyplot(plt)
 
def plot_eda(data):
    if data.empty:
        st.error("No valid data available for EDA.")
        return
    
    st.write("Exploratory Data Analysis:")
    
    # Pairplot
    st.write("Pairplot:")
    sns.pairplot(data, hue=data.columns[-1])
    st.pyplot(plt)
    
    # Heatmap
    st.write("Heatmap:")
    plt.figure(figsize=(10, 8))
    sns.heatmap(data.corr(), annot=True, cmap='coolwarm')
    st.pyplot(plt)
    
    # Boxplot
    st.write("Boxplot:")
    plt.figure(figsize=(10, 6))
    sns.boxplot(data=data)
    st.pyplot(plt)

def main():
    st.title("Application Development for Data Analysis")

    # Load the data
    data = load_data()

    # Display the data
    if data is not None:
        if validate_data(data): 
            st.write("Data Preview:")
            st.dataframe(data.head())

            # Preprocess the data
            data_clean = preprocess_data(data)
            
            # Check if cleaned data is empty
            if data_clean.empty:
                st.error("No valid data available after cleaning. Please upload a dataset with valid numeric values.")
                return
            # Tabs for different visualizations
            tab1, tab2  = st.tabs(["T-SNE", "EDA"])

             with tab1:
                st.header("t-SNE Visualization")
                plot_tsne(data_clean)
                
            with tab2:
                st.header("Exploratory Data Analysis")
                plot_eda(data_clean)
        else:
            st.error("Invalid data structure. Please upload a dataset with the correct format.")

if __name__ == "__main__":
    main()