app.py

import gradio as gr
import pandas as pd
import plotly.graph_objects as go
from sklearn.linear_model import LinearRegression
import numpy as np
from pandas.tseries.offsets import MonthEnd
import datetime

def plot_and_predict(zip, start_year, start_month, prediction_months):
    # Input validation for ZIP code
    if not zip.isdigit() or len(zip) != 5:
        return "Error: Please enter a valid 5-digit ZIP code."
    
    # Input validation for start year
    current_year = datetime.datetime.now().year
    if not start_year.isdigit() or not (2000 <= int(start_year) <= current_year):
        return f"Error: Please enter a valid year between 2000 and {current_year}."

    # Convert start_month to integer and combine year and month into a start date
    try:
        start_month_int = int(start_month)
        start_date = f"{start_year}-{start_month_int:02d}-01"
    except ValueError:
        return "Error: Invalid start month. Please enter a numeric month between 1 and 12."

    start_date = pd.to_datetime(start_date)

    # Read and process the real estate data from Zillow
    df = pd.read_csv('https://files.zillowstatic.com/research/public_csvs/zhvi/Zip_zhvi_uc_sfrcondo_tier_0.33_0.67_sm_sa_month.csv')
    df = df[df['RegionName'] == int(zip)]
    df = df.loc[:, '2000-01-31':]
    df = df.T.reset_index()
    df.columns = ['Date', 'Price']
    df['Date'] = pd.to_datetime(df['Date'])

    # Filter data based on start date
    df = df[df['Date'] >= start_date]

    # Check for empty dataframe
    if df.empty:
        return "Error: No data found for the provided ZIP code or start year/month. Please check your inputs."

    # Train linear regression model
    df['MonthsSinceStart'] = np.arange(len(df))
    X = df['MonthsSinceStart'].values.reshape(-1, 1)
    y = df['Price'].values
    model = LinearRegression()
    model.fit(X, y)

    # Predict future prices
    last_month_index = df['MonthsSinceStart'].iloc[-1]
    future_months = np.array([last_month_index + i for i in range(1, prediction_months + 1)]).reshape(-1, 1)
    predicted_prices = model.predict(future_months)

    # Prepare data for plotting
    historical_prices_trace = go.Scatter(
        x=df['Date'],
        y=df['Price'],
        mode="lines",
        name="Historical Prices"
    )
    future_dates = [df['Date'].iloc[-1] + MonthEnd(i) for i in range(1, prediction_months + 1)]
    predicted_prices_trace = go.Scatter(
        x=future_dates,
        y=predicted_prices,
        mode="lines",
        name="Predicted Prices"
    )
    
    # Plot data
    fig = go.Figure()
    fig.add_trace(historical_prices_trace)
    fig.add_trace(predicted_prices_trace)
    
    fig.update_layout(
        title=f"Real Estate Price Prediction for Zip Code {zip}",
        xaxis_title="Date",
        yaxis_title="Price",
        legend_title_text="Data"
    )

    return fig

# Gradio interface with updated inputs
interface = gr.Interface(
    fn=plot_and_predict,
    inputs=[
        gr.Textbox(label="ZIP Code", placeholder="e.g., 90210"),
        gr.Textbox(label="Start Year", placeholder="e.g., 2020"),
        gr.Dropdown(label="Start Month", choices=[str(i) for i in range(1, 13)]),
        gr.Slider(minimum=1, maximum=60, step=1, label="Prediction Months"),
    ],
    outputs="plot",
    title="Real Estate Price Predictor",
    description="Enter a ZIP code, start year, start month, and the number of months for price prediction."
)

# Launch the app
interface.launch(debug=True)