imtherealbatman/programming_space

1import streamlit as st2import pandas as pd3import numpy as np4import yfinance as yf5import matplotlib.pyplot as plt6import tensorflow as tf7from tensorflow.keras.models import Sequential8from tensorflow.keras.layers import LSTM, Dense, Dropout9from sklearn.preprocessing import MinMaxScaler10 11# Streamlit App Title12st.title("📈 Stock Price Prediction App (LSTM Model)")13 14# Sidebar Inputs15st.sidebar.header("Stock Selection")16ticker = st.sidebar.text_input("Enter Stock Ticker (e.g., AAPL, TSLA, GOOGL)", value="AAPL").upper()17days_to_predict = st.sidebar.slider("Prediction Period (Days)", min_value=5, max_value=60, value=30, step=5)18 19# Load Stock Data20@st.cache_data21def load_data(ticker):22    stock_data = yf.download(ticker, period="5y")23    return stock_data24 25if ticker:26    try:27        df = load_data(ticker)28 29        if df.empty:30            st.error("Invalid Stock Ticker or No Data Available.")31        else:32            # Display Stock Data Overview33            st.write(f"### {ticker} Stock Data Overview")34            st.write(df.tail())35 36            # Plot Historical Stock Prices37            st.write("### 📊 Historical Stock Prices")38            fig, ax = plt.subplots()39            ax.plot(df.index, df['Close'], label='Closing Price', color='blue')40            ax.set_xlabel("Date")41            ax.set_ylabel("Price (USD)")42            ax.legend()43            st.pyplot(fig)44 45            # Prepare Data for LSTM46            scaler = MinMaxScaler(feature_range=(0, 1))47            df_scaled = scaler.fit_transform(df[['Close']])48 49            # Create Sequences for LSTM50            def create_sequences(data, seq_length):51                X, y = [], []52                for i in range(len(data) - seq_length - 1):53                    X.append(data[i:i+seq_length])54                    y.append(data[i+seq_length])55                return np.array(X), np.array(y)56 57            seq_length = 60  # Use last 60 days to predict the next58            X_train, y_train = create_sequences(df_scaled, seq_length)59 60            # Reshape Data for LSTM61            X_train = np.reshape(X_train, (X_train.shape[0], X_train.shape[1], 1))62 63            # Build LSTM Model64            model = Sequential([65                LSTM(50, return_sequences=True, input_shape=(seq_length, 1)),66                Dropout(0.2),67                LSTM(50, return_sequences=False),68                Dropout(0.2),69                Dense(25),70                Dense(1)71            ])72 73            # Compile the Model74            model.compile(optimizer='adam', loss='mean_squared_error')75 76            # Train the Model (Use limited epochs for quick training)77            model.fit(X_train, y_train, epochs=5, batch_size=16, verbose=1)78 79            # Predict Future Prices80            last_days = df_scaled[-seq_length:]  # Get last 60 days81            future_preds = []82 83            for _ in range(days_to_predict):84                input_seq = np.reshape(last_days, (1, seq_length, 1))85                predicted_price = model.predict(input_seq)[0, 0]86                future_preds.append(predicted_price)87                last_days = np.append(last_days[1:], predicted_price).reshape(-1, 1)88 89            # Convert Predictions Back to Original Scale90            future_preds = scaler.inverse_transform(np.array(future_preds).reshape(-1, 1))91 92            # Plot Prediction93            st.write("### 🔮 Stock Price Prediction for Next {} Days".format(days_to_predict))94            future_dates = pd.date_range(df.index[-1], periods=days_to_predict + 1)[1:]95 96            fig2, ax2 = plt.subplots()97            ax2.plot(df.index[-100:], df['Close'].values[-100:], label='Historical Price', color='blue')98            ax2.plot(future_dates, future_preds, label='Predicted Price', color='red', linestyle='dashed')99            ax2.set_xlabel("Date")100            ax2.set_ylabel("Price (USD)")101            ax2.legend()102            st.pyplot(fig2)103 104    except Exception as e:105        st.error(f"Error: {str(e)}")106 107