Update app.py

app.py

@@ -1,37 +1,54 @@
-import pandas as pd
-import numpy as np
-import pickle
 import streamlit as st
+import pickle
+import numpy as np
+import pandas as pd
+
 
-# Load the trained model pipeline (includes preprocessing + classifier)
 with open("weather_classification_model.pkl", "rb") as model_file:
-    pipeline = pickle.load(model_file)  # ✅ This is already a trained model pipeline
+    model = pickle.load(model_file)
+
 
-# Load the Label Encoder
 with open("label_encoder.pkl", "rb") as encoder_file:
     label_encoder = pickle.load(encoder_file)
 
-# Streamlit UI
-st.title("🌦 Weather Type Prediction")
-
-# User Inputs
-Temperature = st.number_input("Enter Temperature (°C)", value=25.0)
-Humidity = st.number_input("Enter Humidity (%)", value=50.0)
-Wind_Speed = st.number_input("Enter Wind Speed (km/h)", value=10.0)
-Precipitation = st.number_input("Enter Precipitation (%)", value=0.0)
-Cloud_Cover = st.selectbox("Select Cloud Cover", ['Clear', 'Partly Cloudy', 'Overcast'])
-UV_Index = st.number_input("Enter UV Index", value=3)
-Season = st.selectbox("Select Season", ['Spring', 'Summer', 'Autumn', 'Winter'])
-Visibility = st.number_input("Enter Visibility (km)", value=10.0)
-Location = st.text_input("Enter Location", "New York")
-
-# Convert input into DataFrame
-input_data = pd.DataFrame([[Temperature, Humidity, Wind_Speed, Precipitation, Cloud_Cover, UV_Index, Season, Visibility, Location]],
-                          columns=['Temperature', 'Humidity', 'Wind Speed', 'Precipitation (%)', 
-                                   'Cloud Cover', 'UV Index', 'Season', 'Visibility (km)', 'Location'])
-
-# Make Prediction
+
+with open("scaler.pkl", "rb") as scaler_file:
+    scaler = pickle.load(scaler_file)
+
+
+st.title("🌦️ Weather Prediction App")
+
+st.write("Enter weather details to predict the weather type.")
+
+
+Temperature = st.number_input("Temperature (°C)", min_value=-50, max_value=50, value=20)
+Humidity = st.number_input("Humidity (%)", min_value=0, max_value=100, value=50)
+Wind_Speed = st.number_input("Wind Speed (km/h)", min_value=0, max_value=150, value=10)
+Precipitation = st.number_input("Precipitation (%)", min_value=0, max_value=100, value=10)
+Cloud_Cover = st.selectbox("Cloud Cover", ["Clear", "Partial", "Overcast"])
+UV_Index = st.number_input("UV Index", min_value=0, max_value=15, value=5)
+Season = st.selectbox("Season", ["Winter", "Spring", "Summer", "Fall"])
+Visibility = st.number_input("Visibility (km)", min_value=0, max_value=50, value=10)
+Location = st.selectbox("Location", ["inland", "mountain", "coastal"])
+
+
+categorical_features = {"Cloud Cover": Cloud_Cover, "Season": Season, "Location": Location}
+for feature, value in categorical_features.items():
+    categorical_features[feature] = label_encoder.transform([value])[0]
+
+
+input_data = np.array([[Temperature, Humidity, Wind_Speed, Precipitation,
+                        categorical_features["Cloud Cover"], UV_Index, 
+                        categorical_features["Season"], Visibility, 
+                        categorical_features["Location"]]])
+
+
+input_data_scaled = scaler.transform(input_data)
+
+
 if st.button("Predict Weather Type"):
-    prediction_numeric = pipeline.predict(input_data)  # ✅ Now uses the trained pipeline
-    predicted_label = label_encoder.inverse_transform(prediction_numeric)  # Convert numeric label to original category
-    st.success(f"🌤️ Predicted Weather Type: **{predicted_label[0]}**")
+    prediction_numeric = model.predict(input_data_scaled)[0]
+    predicted_label = label_encoder.inverse_transform([prediction_numeric])[0]
+    
+    st.success(f"🌤️ Predicted Weather Type: **{predicted_label}**")
+