Update app.py

app.py

@@ -1,59 +1,35 @@
-import streamlit as st
 import numpy as np
 import pandas as pd
+import streamlit as st
+import matplotlib.pyplot as plt
 from sklearn.preprocessing import OneHotEncoder, StandardScaler
 from keras.models import load_model
-import matplotlib.pyplot as plt
-from keras.layers import LSTM
-import keras
-from keras import backend as K
-
-@keras.saving.register_keras_serializable()
-def custom_mse(y_true, y_pred):
-    return K.mean(K.square(y_pred - y_true))
-
-def custom_lstm(*args, **kwargs):
-    kwargs.pop('time_major', None)  # Remove the unsupported argument
-    return LSTM(*args, **kwargs)
-
+from sklearn.metrics import mean_squared_error, mean_absolute_error, r2_score
 
-
-
-# Load the trained model
-try:
-    loaded_model = load_model('solar_irradiance_model.h5', custom_objects={'LSTM': custom_lstm})
-except ValueError as e:
-    st.error(f"Error loading model: {e}")
-
-# Load the dataset
-data = pd.read_csv('Solar_Irradiance.csv')
+# Load the data
+data = pd.read_csv('solar_irradiance.csv')
 data['Latitude'] = data['Latitude'].str.rstrip('°').astype(float)
 data['Longitude'] = data['Longitude'].str.rstrip('°').astype(float)
 
-# One-hot encoder setup
+# Extract features and target
+features = data[['Month', 'Hour', 'Latitude', 'Longitude', 'Panel_Capacity(W)', 'Panel_Efficiency', 'Wind_Speed(km/h)', 'Cloud_Cover(%)', 'temperature (°f)']]
+target = data['Irradiance(W/m^2)']
+
+# Set up the encoder and scaler
 encoder = OneHotEncoder(sparse=False, categories='auto')
-categorical_features = data[['Month', 'Hour']]
+categorical_features = features[['Month', 'Hour']]
 encoder.fit(categorical_features)
-
-# Standard scaler setup
 scaler = StandardScaler()
-numerical_features = data[['Latitude', 'Longitude', 'Panel_Capacity(W)', 'Panel_Efficiency', 'Wind_Speed(km/h)', 'Cloud_Cover(%)', 'temperature (°f)']]
-scaler.fit(numerical_features)
+scaler.fit(features[['Latitude', 'Longitude', 'Panel_Capacity(W)', 'Panel_Efficiency', 'Wind_Speed(km/h)', 'Cloud_Cover(%)', 'temperature (°f)']])
 
-# Function to predict irradiance
-def predict_irradiance(month, hour, latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature):
-    encoded_month_hour = encoder.transform([[month, hour]])
-    scaled_features = scaler.transform([[latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature]])
-    processed_features = np.concatenate((encoded_month_hour, scaled_features), axis=1)
-    reshaped_features = np.reshape(processed_features, (1, 1, processed_features.shape[1]))
-    predicted_irradiance = loaded_model.predict(reshaped_features)
-    return max(predicted_irradiance[0][0], 0.0)
+# Load the saved model
+loaded_model = load_model('solar_irradiance_model.h5')
 
-# Streamlit app UI
-st.title("Solar Irradiance Prediction App")
+# Streamlit app setup
+st.title("Solar Irradiance Prediction")
 st.sidebar.header("Input Parameters")
 
-# User inputs
+# User inputs from sidebar
 month = st.sidebar.selectbox("Month", data['Month'].unique())
 hour = st.sidebar.slider("Hour", 0, 23, 12)
 latitude = st.sidebar.number_input("Latitude", value=28.570633)
@@ -64,21 +40,116 @@ wind_speed = st.sidebar.number_input("Wind Speed (km/h)", value=6.43988)
 cloud_cover = st.sidebar.number_input("Cloud Cover (%)", value=17.7)
 temperature = st.sidebar.number_input("Temperature (°F)", value=55.0)
 
-# Predict and display
-if st.button("Predict"):
-    predicted_irradiance = predict_irradiance(month, hour, latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature)
-    st.write(f"Predicted Solar Irradiance: {predicted_irradiance:.2f} W/m²")
-
-# Visualization
-st.subheader("Actual vs. Predicted Irradiance")
-actual_irradiance = data[data['Month'] == month]['Irradiance(W/m^2)'].values
-predicted_irradiances = [predict_irradiance(month, h, latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature) for h in range(24)]
-
-plt.figure(figsize=(10, 5))
-plt.plot(range(24), actual_irradiance, label='Actual Irradiance')
-plt.plot(range(24), predicted_irradiances, label='Predicted Irradiance')
-plt.xlabel('Hour')
-plt.ylabel('Irradiance (W/m²)')
-plt.title(f'Actual vs. Predicted Irradiance for {month}')
-plt.legend()
-st.pyplot(plt)
+# Function to predict irradiance
+def predict_irradiance(month, hour, latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature):
+    encoded_month_hour = encoder.transform([[month, hour]])
+    scaled_features = scaler.transform([[latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature]])
+    processed_features = np.concatenate((encoded_month_hour, scaled_features), axis=1)
+    reshaped_features = np.reshape(processed_features, (1, 1, processed_features.shape[1]))
+    predicted_irradiance = loaded_model.predict(reshaped_features)
+    return max(predicted_irradiance[0][0], 0.0)
+
+# Display the predicted irradiance
+predicted_irradiance = predict_irradiance(month, hour, latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature)
+st.write(f"Predicted irradiance for {month}, hour {hour}: {predicted_irradiance:.2f} W/m^2")
+
+# Plot actual vs. predicted irradiance
+def plot_actual_vs_predicted(month):
+    actual_irradiance = data[data['Month'] == month]['Irradiance(W/m^2)'].values
+    predicted_irradiances = []
+
+    for hour in range(24):
+        irradiance = predict_irradiance(month, hour, latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature)
+        predicted_irradiances.append(irradiance)
+
+    plt.figure(figsize=(12, 6))
+    plt.plot(range(24), actual_irradiance, label='Actual Irradiance', color='blue')
+    plt.plot(range(24), predicted_irradiances, label='Predicted Irradiance', color='red')
+    plt.xlabel('Hour')
+    plt.ylabel('Irradiance (W/m^2)')
+    plt.title(f'Actual vs. Predicted Irradiance for {month}')
+    plt.legend()
+    st.pyplot(plt)
+
+# Display the plot
+plot_actual_vs_predicted(month)
+
+# Function to calculate and display evaluation metrics
+def display_evaluation_metrics():
+    metrics_df = pd.DataFrame(columns=['Month', 'MSE', 'RMSE', 'MAE', 'R-squared'])
+
+    for month in data['Month'].unique():
+        actual_irradiance = get_actual_irradiance(month)
+        predicted_irradiance = []
+
+        for hour in range(24):
+            irradiance = predict_irradiance(month, hour, latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature)
+            predicted_irradiance.append(irradiance)
+
+        mse = mean_squared_error(actual_irradiance, predicted_irradiance)
+        rmse = np.sqrt(mse)
+        mae = mean_absolute_error(actual_irradiance, predicted_irradiance)
+        r2 = r2_score(actual_irradiance, predicted_irradiance)
+
+        metrics_df = metrics_df.append({'Month': month, 'MSE': mse, 'RMSE': rmse, 'MAE': mae, 'R-squared': r2}, ignore_index=True)
+
+    st.write("Evaluation Metrics for Each Month")
+    st.dataframe(metrics_df)
+
+# Display evaluation metrics
+display_evaluation_metrics()
+
+# Function to plot actual vs. predicted irradiance scatter plot
+def plot_irradiance_scatter(month):
+    actual_irradiance = get_actual_irradiance(month)
+    predicted_irradiances = []
+
+    for hour in range(24):
+        irradiance = predict_irradiance(month, hour, latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature)
+        predicted_irradiances.append(irradiance)
+
+    plt.figure(figsize=(8, 6))
+    plt.scatter(range(24), actual_irradiance, label='Actual Irradiance', color='blue')
+    plt.scatter(range(24), predicted_irradiances, label='Predicted Irradiance', color='red')
+    plt.xlabel('Hour')
+    plt.ylabel('Irradiance (W/m^2)')
+    plt.title(f'Actual vs. Predicted Irradiance for {month}')
+    plt.legend()
+    st.pyplot(plt)
+
+# Example usage: scatter plot for selected month
+plot_irradiance_scatter(month)
+
+# Function to plot hour vs. irradiance for all months
+def plot_hour_vs_irradiance():
+    months = data['Month'].unique()
+    hour_range = range(24)
+
+    predicted_irradiances = np.zeros((len(months), 24))
+    actual_irradiances = np.zeros((len(months), 24))
+
+    for i, month in enumerate(months):
+        for hour in hour_range:
+            if hour in range(6) or hour in range(18, 24):
+                irradiance = 0  # Set predicted irradiance to 0
+            else:
+                irradiance = predict_irradiance(month, hour, latitude, longitude, panel_capacity, panel_efficiency, wind_speed, cloud_cover, temperature)
+            predicted_irradiances[i][hour] = irradiance
+            actual_irradiances[i][hour] = get_actual_irradiance(month)[hour]
+
+    bar_width = 0.35
+    index = np.arange(len(hour_range))
+
+    plt.figure(figsize=(12, 6))
+    plt.bar(index, predicted_irradiances.mean(axis=0), bar_width, label='Predicted Irradiance')
+    plt.bar(index + bar_width, actual_irradiances.mean(axis=0), bar_width, label='Actual Irradiance')
+
+    plt.xlabel('Hour')
+    plt.ylabel('Irradiance (W/m^2)')
+    plt.title('Hour vs. Irradiance (Average for All Months)')
+    plt.xticks(index + bar_width/2, hour_range)
+    plt.legend()
+    st.pyplot(plt)
+
+# Plot hour vs. irradiance for all months
+plot_hour_vs_irradiance()