diff --git a/app.py b/app.py
index ce765c6f11f76d8d8610c0ef0a0642cc3d96ee6d..0510bee10f750920df06d788c0c58005402aaebc 100644
--- a/app.py
+++ b/app.py
@@ -6,47 +6,47 @@ from sklearn.preprocessing import LabelEncoder
import joblib # For loading Scikit-Learn models
# Load the dataset
-try:
- df = pd.read_csv("FakeDataset.csv")
-except FileNotFoundError:
- df = None
- st.error("Dataset not found. Please ensure 'FakeDataset.csv' is in the correct path.")
+df = pd.read_csv("FakeDataset.csv")
# Load Scikit-Learn models
-try:
- model_1 = joblib.load("model_1.pkl") # Model for satisfaction prediction
- model_2 = joblib.load("model_2.pkl") # Model for flight delay prediction
-except FileNotFoundError:
- st.error("Model files not found. Please ensure 'model_1.pkl' and 'model_2.pkl' are in the correct path.")
+model_1 = joblib.load("model_1.pkl") # Model for satisfaction prediction
+model_2 = joblib.load("model_2.pkl") # Model for flight delay prediction
# Dynamically create the encoders if they don't exist
country_code_encoder = None
airport_encoder = None
-if df is not None:
- if not os.path.exists("country_code_encoder.pkl"):
- st.warning("Country code encoder file not found. Creating a new encoder dynamically...")
- if "Airport Country Code" in df.columns:
- country_code_encoder = LabelEncoder()
- df["Airport Country Code"] = country_code_encoder.fit_transform(df["Airport Country Code"])
- joblib.dump(country_code_encoder, "country_code_encoder.pkl")
- st.success("Country code encoder created and saved as 'country_code_encoder.pkl'.")
- else:
- st.error("The dataset is missing the 'Airport Country Code' column.")
+if not os.path.exists("country_code_encoder.pkl"):
+ st.warning("Country code encoder file not found. Creating a new encoder dynamically...")
+ if "Airport Country Code" in df.columns:
+ # Create a new encoder using the dataset
+ country_code_encoder = LabelEncoder()
+ df["Airport Country Code"] = country_code_encoder.fit_transform(df["Airport Country Code"])
+
+ # Save the encoder for future use
+ joblib.dump(country_code_encoder, "country_code_encoder.pkl")
+ st.success("Country code encoder has been created and saved as 'country_code_encoder.pkl'.")
else:
- country_code_encoder = joblib.load("country_code_encoder.pkl")
-
- if not os.path.exists("airport_encoder.pkl"):
- st.warning("Airport encoder file not found. Creating a new encoder dynamically...")
- if "Arrival Airport" in df.columns:
- airport_encoder = LabelEncoder()
- df["Arrival Airport"] = airport_encoder.fit_transform(df["Arrival Airport"])
- joblib.dump(airport_encoder, "airport_encoder.pkl")
- st.success("Airport encoder created and saved as 'airport_encoder.pkl'.")
- else:
- st.error("The dataset is missing the 'Arrival Airport' column.")
+ st.error("The dataset is missing the 'Airport Country Code' column.")
+else:
+ # Load the encoder if it already exists
+ country_code_encoder = joblib.load("country_code_encoder.pkl")
+
+if not os.path.exists("airport_encoder.pkl"):
+ st.warning("Airport encoder file not found. Creating a new encoder dynamically...")
+ if "Arrival Airport" in df.columns:
+ # Create a new encoder using the dataset
+ airport_encoder = LabelEncoder()
+ df["Arrival Airport"] = airport_encoder.fit_transform(df["Arrival Airport"])
+
+ # Save the encoder for future use
+ joblib.dump(airport_encoder, "airport_encoder.pkl")
+ st.success("Airport encoder has been created and saved as 'airport_encoder.pkl'.")
else:
- airport_encoder = joblib.load("airport_encoder.pkl")
+ st.error("The dataset is missing the 'Arrival Airport' column.")
+else:
+ # Load the encoder if it already exists
+ airport_encoder = joblib.load("airport_encoder.pkl")
# Initialize Streamlit
st.title("Airline Chatbot & Data Explorer with Predictions")
@@ -57,19 +57,114 @@ st.image("Chatbot_image.gif", caption="Welcome to the Airline Chatbot & Data Exp
# Add navigation
menu = st.sidebar.selectbox("Select a Feature", ["Chatbot", "Data Explorer", "Visualizations", "Predictions"])
-# Functions for different sections
+def show_chatbot_section():
+ st.subheader("Chat with Airline Assistant")
+
+ # Initialize session states for conversation flow
+ if "greeted" not in st.session_state:
+ st.session_state.greeted = False
+ if "help_query" not in st.session_state:
+ st.session_state.help_query = False
+ if "details_fetched" not in st.session_state:
+ st.session_state.details_fetched = False
+ if "farewell" not in st.session_state:
+ st.session_state.farewell = False
+
+ # Reset button functionality
+ if st.button("Reset Chatbot"):
+ st.session_state.greeted = False
+ st.session_state.help_query = False
+ st.session_state.details_fetched = False
+ st.session_state.farewell = False
+ st.success("The chatbot has been reset. How can I help you?")
+
+ # Greeting and initial input
+ if not st.session_state.greeted:
+ user_input = st.text_input("Start by saying 'Hey', 'Good Morning', or 'Hello':")
+ if user_input:
+ if any(greet in user_input.lower() for greet in ["hey", "good morning", "hello", "hi", "hey there", "good afternoon", "good evening", "greetings", "good day", "howdy", "sup", "what's up"]):
+ st.write("Hello! How can I assist you today?")
+ st.session_state.greeted = True
+ else:
+ st.warning("Please start with a greeting like 'Hey' or 'Good Morning'.")
+
+ # Respond to user query
+ if st.session_state.greeted and not st.session_state.help_query:
+ user_query = st.text_input("What would you like assistance with? (e.g., 'I want to know my flight details'):")
+ if user_query:
+ if any(keyword in user_query.lower() for keyword in ["details", "my details", "know my details", "flight details", "flight status", "my flight details", "get details", "get my details", "flight information", "flight info"]):
+ st.success("Understood, let's proceed.")
+ st.session_state.help_query = True
+ else:
+ st.info("Let me know how I can assist you further.")
+
+ # Passenger ID and Last Name functionality
+ if st.session_state.help_query and not st.session_state.details_fetched:
+ passenger_id = st.text_input("Please enter your Passenger ID:")
+ last_name = st.text_input("Please enter your Last Name:")
+
+ if st.button("Get Flight Details"):
+ if not passenger_id:
+ st.warning("Please enter your Passenger ID.")
+ elif not last_name:
+ st.warning("Please enter your Last Name.")
+ else:
+ result = df[df.applymap(lambda x: passenger_id.lower() in str(x).lower() if pd.notnull(x) else False).any(axis=1)]
+ result = result[result["Last Name"].str.lower() == last_name.lower()]
+ if not result.empty:
+ st.success("**Here are your flight details:**")
+ for index, row in result.iterrows():
+ st.markdown(f"**Passenger ID:** <code>{row['Passenger ID']}</code>", unsafe_allow_html=True)
+ st.markdown(f"**First Name:** <code>{row['First Name']}</code>", unsafe_allow_html=True)
+ st.markdown(f"**Last Name:** <code>{row['Last Name']}</code>", unsafe_allow_html=True)
+ st.markdown(f"**Gender:** <code>{row['Gender']}</code>", unsafe_allow_html=True)
+ st.markdown(f"**Age:** <code>{row['Age']}</code>", unsafe_allow_html=True)
+ st.markdown(f"**Nationality:** <code>{row['Nationality']}</code>", unsafe_allow_html=True)
+ st.markdown(f"**Airport Name:** <code>{row['Airport Name']}</code>", unsafe_allow_html=True)
+ st.markdown(f"**Departure Date:** <code>{row['Departure Date']}</code>", unsafe_allow_html=True)
+ st.markdown(f"**Arrival Airport:** <code>{row['Arrival Airport']}</code>", unsafe_allow_html=True)
+ st.markdown(f"**Pilot Name:** <code>{row['Pilot Name']}</code>", unsafe_allow_html=True)
+ st.markdown(f"**Flight Status:** <code>{row['Flight Status']}</code>", unsafe_allow_html=True)
+ st.write("---")
+ st.session_state.details_fetched = True
+ else:
+ st.error("No matching records found. Please check the Passenger ID or Last Name.")
+
+ # Further assistance or goodbye
+ if st.session_state.details_fetched and not st.session_state.farewell:
+ further_assistance = st.text_input("Do you need any further assistance? (Yes/No)")
+ if further_assistance:
+ if "no" in further_assistance.lower() or "goodbye" in further_assistance.lower():
+ st.success("Happy to assist you!")
+ st.session_state.farewell = True
+ elif "yes" in further_assistance.lower():
+ st.info("Please describe your next query or ask for more details.")
+ else:
+ st.warning("Please respond with 'Yes' or 'No'.")
+
+ # Reset message after farewell
+ if st.session_state.farewell:
+ st.info("Press the **Reset Chatbot** button to restart the conversation and ask for other details.")
+
+
+def show_data_explorer_section():
+ st.subheader("Explore the Airline Dataset")
+
+ # Display dataset columns
+ columns = st.multiselect("Select columns to view:", options=df.columns.tolist(), default=df.columns.tolist())
+ st.dataframe(df[columns])
+
+ # Filter by Flight Status
+ flight_statuses = st.multiselect("Filter by Flight Status:", options=df["Flight Status"].unique().tolist(), default=df["Flight Status"].unique().tolist())
+ filtered_data = df[df["Flight Status"].isin(flight_statuses)]
+ st.write(f"Filtered Records: {len(filtered_data)}")
+ st.dataframe(filtered_data)
+
+
+# Updated Visualization Function
def show_visualizations_section(df):
st.subheader("Visualize the Airline Dataset")
- # Flight Status Bar Chart
- status_counts = df["Flight Status"].value_counts().reset_index()
- status_counts.columns = ["Flight Status", "Count"]
- status_chart = alt.Chart(status_counts).mark_bar().encode(
- x=alt.X("Flight Status", title="Flight Status"),
- y=alt.Y("Count", title="Count"),
- tooltip=["Flight Status", "Count"]
- ).properties(title="Flight Status Distribution")
- st.altair_chart(status_chart, use_container_width=True)
-
+
# Passenger Nationality Pie Chart
nationality_counts = df["Nationality"].value_counts().reset_index().head(10)
nationality_counts.columns = ["Nationality", "Count"]
@@ -88,15 +183,115 @@ def show_visualizations_section(df):
).properties(title="Passenger Age Distribution")
st.altair_chart(age_histogram, use_container_width=True)
+ # Flight Status Bar Chart
+ status_counts = df["Flight Status"].value_counts().reset_index()
+ status_counts.columns = ["Flight Status", "Count"]
+ status_chart = alt.Chart(status_counts).mark_bar().encode(
+ x=alt.X("Flight Status", title="Flight Status"),
+ y=alt.Y("Count", title="Count"),
+ tooltip=["Flight Status", "Count"]
+ ).properties(title="Flight Status Distribution")
+ st.altair_chart(status_chart, use_container_width=True)
+
+def show_predictions_section():
+ st.subheader("Make Predictions with Scikit-Learn Models")
+
+ # Input fields for Model 1 (Passenger Satisfaction Prediction)
+ st.write("### Passenger Satisfaction Prediction")
+ age = st.slider("Age", min_value=18, max_value=100, value=30)
+ gender = st.selectbox("Gender", ["Male", "Female"])
+ flight_class = st.selectbox("Flight Class", ["Economy", "Business", "First Class"])
+ flight_duration = st.number_input("Flight Duration (hours)", min_value=0.0, value=2.0)
+
+ # Map gender and flight class to numerical values for model input
+ gender_map = {"Male": 0, "Female": 1}
+ flight_class_map = {"Economy": 0, "Business": 1, "First Class": 2}
+
+ gender_num = gender_map[gender]
+ flight_class_num = flight_class_map[flight_class]
+
+ # Model 1 Prediction
+ if st.button("Predict Satisfaction"):
+ satisfaction_input = pd.DataFrame(
+ [[age, gender_num, flight_class_num, flight_duration]],
+ columns=["Age", "Gender", "Flight Class", "Flight Duration"]
+ )
+ satisfaction_prediction = model_1.predict(satisfaction_input)
+ st.write(f"Predicted Passenger Satisfaction: {'Satisfied' if satisfaction_prediction[0] == 1 else 'Not Satisfied'}")
+
+ # Generate live graph for Satisfaction Prediction
+ satisfaction_data = pd.DataFrame({
+ "Category": ["Not Satisfied", "Satisfied"],
+ "Prediction": [1 - satisfaction_prediction[0], satisfaction_prediction[0]],
+ })
+ satisfaction_chart = alt.Chart(satisfaction_data).mark_bar().encode(
+ x="Category",
+ y="Prediction",
+ color="Category"
+ ).properties(title="Passenger Satisfaction Prediction")
+ st.altair_chart(satisfaction_chart, use_container_width=True)
+
+ st.write("---") # Separator
+
+ # Input fields for Model 2 (Flight Delay Prediction)
+ st.write("### Flight Delay Prediction")
+ airport_country_code = st.text_input("Airport Country Code (e.g., US)").upper()
+ arrival_airport = st.text_input("Arrival Airport Code (e.g., JFK)").upper()
+ departure_time = st.number_input("Departure Time (24-hour format, e.g., 13.5 for 1:30 PM)", min_value=0.0, max_value=24.0)
+
+ if not country_code_encoder or not airport_encoder:
+ st.error("Encoders are not available. Predictions cannot proceed.")
+ return
+
+ if st.button("Predict Delay"):
+ try:
+ # Encode inputs
+ if airport_country_code not in country_code_encoder.classes_:
+ st.error(f"Invalid Airport Country Code: {airport_country_code}")
+ return
+
+ if arrival_airport not in airport_encoder.classes_:
+ st.error(f"Invalid Arrival Airport Code: {arrival_airport}")
+ return
+
+ airport_country_code_encoded = country_code_encoder.transform([airport_country_code])[0]
+ arrival_airport_encoded = airport_encoder.transform([arrival_airport])[0]
+
+ # Prepare input for prediction as a DataFrame
+ delay_input = pd.DataFrame(
+ [[airport_country_code_encoded, arrival_airport_encoded, departure_time]],
+ columns=["Airport Country Code", "Arrival Airport", "Departure Time"]
+ )
+
+ # Perform prediction
+ delay_prediction = model_2.predict(delay_input)
+
+ st.write(f"Predicted Flight Delay: {'Delayed' if delay_prediction[0] == 1 else 'On Time'}")
+
+ # Generate live graph for Delay Prediction
+ delay_data = pd.DataFrame({
+ "Category": ["On Time", "Delayed"],
+ "Prediction": [1 - delay_prediction[0], delay_prediction[0]],
+ })
+ delay_chart = alt.Chart(delay_data).mark_bar().encode(
+ x="Category",
+ y="Prediction",
+ color="Category"
+ ).properties(title="Flight Delay Prediction")
+ st.altair_chart(delay_chart, use_container_width=True)
+ except ValueError as e:
+ st.error(f"Error: {e}. Please ensure the input values are valid and match the training data.")
+
# Main Navigation
if menu == "Chatbot":
- st.subheader("Chatbot functionality to be implemented...")
+ show_chatbot_section()
elif menu == "Data Explorer":
- st.subheader("Data Explorer functionality to be implemented...")
+ show_data_explorer_section()
elif menu == "Visualizations":
if df is not None:
- show_visualizations_section(df)
+ show_visualizations_section(df) # Pass the dataframe as an argument
else:
st.error("Dataset not loaded. Please check the file path.")
elif menu == "Predictions":
- st.subheader("Prediction functionality to be implemented...")
+ show_predictions_section()
+