app.py

import streamlit as st
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder, StandardScaler
from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.metrics import confusion_matrix, classification_report
import matplotlib.pyplot as plt
import seaborn as sns
import re

st.title("Expense Category Prediction")

# Load data from CSV
df = pd.read_csv("financial_data.csv", sep='\s\s+', engine='python')

# Data Preprocessing
def preprocess_data(df):

    # Clean the date column
    df['Date'] = df['Date'].str.extract(r'(\d{4}-\d{2}-\d{2})')
    
    # Forward fill missing dates
    df['Date'] = df['Date'].ffill()
    
    # Remove rows with missing dates
    df.dropna(subset=['Date'], inplace=True)
    
    # Convert 'Date' to datetime objects
    df['Date'] = pd.to_datetime(df['Date'])

    # Fill missing values in 'Expense_Category' and 'Description' with 'Unknown'
    df['Expense_Category'] = df['Expense_Category'].fillna('Unknown')
    df['Description'] = df['Description'].fillna('Unknown')

    # Convert 'Amount' to numeric, fill missing with 0
    df['Amount'] = pd.to_numeric(df['Amount'], errors='coerce').fillna(0)
    
    # Date Feature Engineering
    df['Month'] = df['Date'].dt.month
    df['DayOfWeek'] = df['Date'].dt.dayofweek

    # Description Text Processing
    def clean_text(text):
        text = text.lower()
        text = re.sub(r'[^a-zA-Z0-9\s]', '', text)
        return text

    df['Description_Cleaned'] = df['Description'].apply(clean_text)

    # TF-IDF Vectorization
    tfidf_vectorizer = TfidfVectorizer(max_features=100)  # Limiting features for simplicity
    tfidf_features = tfidf_vectorizer.fit_transform(df['Description_Cleaned']).toarray()
    tfidf_df = pd.DataFrame(tfidf_features, index=df.index)

    # Combine Features
    features_df = pd.concat([df[['Amount', 'Month', 'DayOfWeek']], tfidf_df], axis=1)

    # Encode the target variable
    label_encoder = LabelEncoder()
    df['Expense_Category_Encoded'] = label_encoder.fit_transform(df['Expense_Category'])

    # Select features and target
    X = features_df
    y = df['Expense_Category_Encoded']

    # Scale the features
    scaler = StandardScaler()
    X = scaler.fit_transform(X)

    return X, y, label_encoder, df  # Return the original dataframe

X, y, label_encoder, df = preprocess_data(df.copy())

# Split data
X_train, X_test, y_train, y_test = train_test_split(
    X, y, test_size=0.2, random_state=42)

# --- Models ---
models = {
    "Random Forest": RandomForestClassifier(random_state=42),
    "Gradient Boosting": GradientBoostingClassifier(random_state=42)
}

# --- Streamlit Tabs ---
tabs = st.tabs(list(models.keys()))

for tab, model_name in zip(tabs, models.keys()):
    with tab:
        st.header(model_name)
        model = models[model_name]
        model.fit(X_train, y_train)
        y_pred = model.predict(X_test)

        # --- Confusion Matrix ---
        st.subheader("Confusion Matrix")
        cm = confusion_matrix(y_test, y_pred)
        plt.figure(figsize=(8, 6))
        sns.heatmap(cm, annot=True, fmt="d", cmap="Blues")
        plt.xlabel("Predicted")
        plt.ylabel("Actual")
        st.pyplot(plt.gcf())

        # --- Classification Report ---
        st.subheader("Classification Report")
        cr = classification_report(y_test, y_pred,
                                   target_names=label_encoder.inverse_transform(
                                       df['Expense_Category_Encoded'].unique()),
                                   zero_division=0)  # Get original category names
        st.text(cr)

        # --- Remarks ---
        st.subheader("Remarks")
        st.write("Model Performance Analysis:")
        st.write(
            f"The {model_name} model's performance in predicting Expense Categories is shown above.")
        st.write("Key Metrics:")
        st.write(
            "-   The model uses a combination of expense amount, time-based features, and text descriptions to predict the expense category."
        )
        st.write(
            "-   The classification report provides insights into the model's precision, recall, and F1-score for each expense category."
        )