app.py

import pandas as pd
import streamlit as st
import numpy as np
import sklearn
import joblib
import plotly.express as px


pipeline = joblib.load('toolkit/Pipeline.joblib')
model = joblib.load('toolkit/model.joblib')

df = pd.read_csv('Dataset/predicted_data_access_consumption_2015_2029.csv')


st.set_page_config(
    page_title='Electricity Access Prediction App',
    page_icon=':cityscape:'
)



# Add a title and subtitle
st.write("<center><h1>Electricity Access Prediction App ⚡</h1></center>", unsafe_allow_html=True)

# Sidebar navigation
page = st.sidebar.selectbox("Select Page", ["Predicted Dataset", "Prediction"])



if page == "Predicted Dataset":
    # Add a section for displaying the preloaded DataFrame
    st.subheader("Predicted Dataset")

    # Display the DataFrame
    st.write("Actual and predicted values of our dataset (2015-2029) from the IMF:")
    st.dataframe(df) 

    # Add a selectbox to choose a country
    country = st.selectbox("Select a Country", df['Country'].unique())

    # Filter the DataFrame based on the selected country
    filtered_df = df[df['Country'] == country]

    # Display the filtered DataFrame
    st.write(f"Data for {country}:")
    st.dataframe(filtered_df)

    st.subheader(f"Access to Electricity in {country} Over Time")
    fig = px.line(filtered_df, x='Year', y='Predicted_Access_to_electricity', title=f'Access to Electricity in {country} Over Time', markers=True)
    st.plotly_chart(fig)





elif page == "Prediction":
    
    st.write("This app uses machine learning to predict access to electricity based on certain input parameters. Simply enter the required information and click 'Predict' to get a prediction!")
    st.subheader("Enter the details to predict access to electricity by % of population")

    st.sidebar.write("""
    ## Data Input Explanation
    This app predicts access to electricity based on the following input parameters:
    - **Year**: The year for which the prediction is being made.
    - **Income Group**: The income classification of the country (Low income, Lower middle income, Upper middle income, High income).
    - **Population**: The total population of the country.
    - **GDP per capita (USD)**: The Gross Domestic Product per capita in current US dollars.
    - **Inflation (annual %)**: The annual inflation rate.
    - **Consumption (kWh per capita)**: The electricity consumption per capita in kilowatt-hours.
    """)

    # Set up the layout
    col1, col2 = st.columns(2)

    # Create the input fields
    input_data = {}
    with col1:
        input_data['Year'] = st.number_input("Input Year", min_value=1990, max_value=2030, step=1)
        input_data['IncomeGroup'] = st.radio("Pick an income group", ["Lower middle income", "Upper middle income", "High income", "Low income"])
        input_data['Population'] = st.slider("Population of the country", min_value=9000, max_value=10000000000)

    with col2:
        input_data['GDP_per_capita_USD'] = st.slider("Enter GDP per capita (Current $)", min_value=20.0, max_value=250000.0, step=0.1)
        input_data['Inflation_annual_percent'] = st.slider("Enter Consumer Price Index", min_value=-20.0, max_value=25000.0, step=0.1)
        input_data['Consumption (kWh per capita)'] = st.slider("Enter Electricity Consumption", min_value=10.0, max_value=55000.0, step=0.1)

    # Define the custom CSS
    predict_button_css = """
        <style>
        .predict-button {
            background-color: #C4C4C4;
            color: gray;
            padding: 0.75rem 2rem;
            border-radius: 0.5rem;
            border: none;
            font-size: 1.1rem;
            font-weight: bold;
            text-align: center;
            margin-top: 2rem;
        }
        </style>
    """

    # Display the custom CSS
    st.markdown(predict_button_css, unsafe_allow_html=True)

    # Create a button to make a prediction
    if st.button("Predict", key="predict_button", help="Click to make a prediction."):
        # Convert the input data to a pandas DataFrame
        input_df = pd.DataFrame([input_data])

        # Selecting categorical and numerical columns separately
        numerical = input_df.select_dtypes(include=[np.number]).columns.tolist()
        categorical = input_df.select_dtypes(exclude=[np.number]).columns.tolist()

        # Fit and transform the data
        X_processed = pipeline.transform(input_df)

        # Extracting feature names for numerical columns
        num_feature_names = numerical

        # Extracting feature names for categorical columns after one-hot encoding
        cat_encoder = pipeline.named_steps['preprocessor'].named_transformers_['cat'].named_steps['onehot']
        cat_feature_names = cat_encoder.get_feature_names_out(categorical)

        # Concatenating numerical and categorical feature names
        feature_names = num_feature_names + list(cat_feature_names)

        # Convert X_processed to DataFrame
        final_df = pd.DataFrame(X_processed, columns=feature_names)

        # Make a prediction
        prediction = model.predict(final_df)[0]

        # Display the prediction
        st.write(f"The predicted Access to electricity is: {prediction}.")
        st.table(input_df)