Code Editor for app.py

import streamlit as st
import pandas as pd
import numpy as np
import plotly.express as px
import plotly.graph_objects as go
from st_aggrid import AgGrid, GridOptionsBuilder
from datetime import datetime, timedelta

st.title("Forecast Trend Analyzer")

# Create sample data if none exists in session state
if 'data' not in st.session_state:
    dates = pd.date_range(start='2023-01-01', end='2023-12-31', freq='M')
    values = np.random.normal(100, 10, len(dates))
    st.session_state.data = pd.DataFrame({
        'Date': dates,
        'Value': values
    })

# Grid options
gb = GridOptionsBuilder.from_dataframe(st.session_state.data)
gb.configure_default_column(editable=True)
gb.configure_column('Date', editable=False)
grid_options = gb.build()

# Sidebar controls
st.sidebar.header("Forecast Settings")
forecast_periods = st.sidebar.slider("Forecast Periods (Months)", 1, 12, 6)
confidence_interval = st.sidebar.slider("Confidence Interval (%)", 70, 95, 85)

# Display grid
st.subheader("Historical Data")
grid_response = AgGrid(
    st.session_state.data,
    gridOptions=grid_options,
    height=300,
    fit_columns_on_grid_load=True
)

# Get updated data from grid
df = grid_response['data']

# Calculate forecast
def calculate_forecast(data, periods, confidence):
    X = np.arange(len(data))
    y = data['Value'].values
    
    # Fit linear regression
    coefficients = np.polyfit(X, y, 1)
    trend = np.poly1d(coefficients)
    
    # Generate forecast
    forecast_X = np.arange(len(data), len(data) + periods)
    forecast_y = trend(forecast_X)
    
    # Calculate confidence intervals
    std_dev = np.std(y - trend(X))
    z_score = {70: 1.04, 75: 1.15, 80: 1.28, 85: 1.44, 90: 1.645, 95: 1.96}
    margin = z_score[confidence] * std_dev
    
    return forecast_y, margin

# Generate forecast
forecast_y, margin = calculate_forecast(df, forecast_periods, confidence_interval)

# Create forecast dates
last_date = pd.to_datetime(df['Date'].iloc[-1])
forecast_dates = pd.date_range(start=last_date + timedelta(days=31), 
                             periods=forecast_periods, 
                             freq='M')

# Create plot
fig = go.Figure()

# Add historical data
fig.add_trace(go.Scatter(
    x=df['Date'],
    y=df['Value'],
    name='Historical Data',
    mode='lines+markers'
))

# Add forecast
fig.add_trace(go.Scatter(
    x=forecast_dates,
    y=forecast_y,
    name='Forecast',
    mode='lines',
    line=dict(dash='dash')
))

# Add confidence intervals
fig.add_trace(go.Scatter(
    x=forecast_dates,
    y=forecast_y + margin,
    fill=None,
    mode='lines',
    line=dict(width=0),
    showlegend=False
))

fig.add_trace(go.Scatter(
    x=forecast_dates,
    y=forecast_y - margin,
    fill='tonexty',
    mode='lines',
    line=dict(width=0),
    name=f'{confidence_interval}% Confidence Interval'
))

# Update layout
fig.update_layout(
    title='Historical Data and Forecast',
    xaxis_title='Date',
    yaxis_title='Value',
    hovermode='x unified'
)

# Display plot
st.plotly_chart(fig, use_container_width=True)

# Display forecast summary
st.subheader("Forecast Summary")
forecast_df = pd.DataFrame({
    'Date': forecast_dates,
    'Forecast': forecast_y,
    'Lower Bound': forecast_y - margin,
    'Upper Bound': forecast_y + margin
}).round(2)

st.dataframe(forecast_df)