app.py

import yfinance as yf
import numpy as np
import pandas as pd
import plotly.graph_objs as go
import gradio as gr

def plot_volatility_bands(ticker, reference_year):
    
    # Retrieving historical data and performing some preprocessing
    df = yf.download(ticker)
    df['Returns'] = df['Adj Close'].pct_change(1)
    df['Adj Low'] = df['Low'] - (df['Close'] - df['Adj Close'])
    df['Adj High'] = df['High'] - (df['Close'] - df['Adj Close'])
    df['Adj Open'] = df['Open'] - (df['Close'] - df['Adj Close'])
    df = df.fillna(0)
    
    # Obtaining the annualized volatility
    T = 20
    df['Annualized_Vol'] = np.round(df['Returns'].rolling(T).std()*np.sqrt(252), 2)
    
    # Calculating Bands
    High_Band_1std = df.loc[reference_year]["Annualized_Vol"][-1]*df.loc[reference_year]["Adj Close"][-1] + df.loc[reference_year]["Adj Close"][-1]
    Low_Band_1std = df.loc[reference_year]["Adj Close"][-1] - df.loc[reference_year]["Annualized_Vol"][-1]*df.loc[reference_year]["Adj Close"][-1]
    High_Band_2std =  2*df.loc[reference_year]["Annualized_Vol"][-1]*df.loc[reference_year]["Adj Close"][-1] + df.loc[reference_year]["Adj Close"][-1]
    Low_Band_2std =  df.loc[reference_year]["Adj Close"][-1] - 2*df.loc[reference_year]["Annualized_Vol"][-1]*df.loc[reference_year]["Adj Close"][-1]
    High_Band_3std =  3*df.loc[reference_year]["Annualized_Vol"][-1]*df.loc[reference_year]["Adj Close"][-1] + df.loc[reference_year]["Adj Close"][-1]
    Low_Band_3std =  df.loc[reference_year]["Adj Close"][-1] - 3*df.loc[reference_year]["Annualized_Vol"][-1]*df.loc[reference_year]["Adj Close"][-1]

    
    # Creating Candlestick chart
    candlestick = go.Candlestick(x = df.loc[str(int(reference_year) + 1)].index,
                             open = df.loc[str(int(reference_year) + 1)]['Adj Open'],
                             high = df.loc[str(int(reference_year) + 1)]['Adj High'],
                             low = df.loc[str(int(reference_year) + 1)]['Adj Low'],
                             close = df.loc[str(int(reference_year) + 1)]['Adj Close'],
                             increasing = dict(line=dict(color = 'red')),
                             decreasing = dict(line=dict(color = 'black')),
                             name = 'Candlesticks')


    # Defining layout
    layout = go.Layout(title = {'text': f'<b>Volatility-Based Supply and Demand Levels ({ticker})<br><br><sup>&nbsp;&nbsp;&nbsp;&nbsp;<i>Yearly Forecast - {str(int(reference_year) + 1)}</i></sup></b>',
                                'x': .035, 'xanchor': 'left'},
                       yaxis = dict(title = '<b>Price (USD)</b>',
                                   tickfont=dict(size=16)),
                       xaxis = dict(title = '<b>Date</b>'),
                       template = 'seaborn',
                       plot_bgcolor = '#F6F5F5',
                       paper_bgcolor = '#F6F5F5',
                       height = 450, width = 650,
                       showlegend=False,
                       xaxis_rangeslider_visible = False)

    fig = go.Figure(data = [candlestick], layout = layout)
    
    # Fixing the empty spaces in the X-Axis
    dt_all = pd.date_range(start = df.index[0]
                           , end = df.index[-1]
                           , freq = "D")
    dt_all_py = [d.to_pydatetime() for d in dt_all]
    dt_obs_py = [d.to_pydatetime() for d in df.index]

    dt_breaks = [d for d in dt_all_py if d not in dt_obs_py]

    fig.update_xaxes(
        rangebreaks = [dict(values = dt_breaks)]
    )

    # 1σ
    fig.add_hline(y = High_Band_1std, line_width = 2, line_dash = "dot", line_color = "green")
    fig.add_hline(y = Low_Band_1std, line_width = 2, line_dash = "dot", line_color = "red")

    # 2σ
    fig.add_hline(y = High_Band_2std, line_width = 4, line_dash = "dash", line_color = "green")
    fig.add_hline(y = Low_Band_2std, line_width = 4, line_dash = "dash", line_color = "red")

    # 3σ
    fig.add_hline(y = High_Band_3std, line_width = 6, line_dash = "dashdot", line_color = "green")
    fig.add_hline(y = Low_Band_3std, line_width = 6, line_dash = "dashdot", line_color = "red")

    fig.show()
    
    # Printing Supply and Demand Levels
    print(f"\nVolatility-Based Supply and Demand Levels for {ticker} in {int(reference_year) + 1}\n")
    print(f"Supply Level 3σ: {High_Band_3std.round(2)}\n")
    print(f"Supply Level 2σ: {High_Band_2std.round(2)}\n")
    print(f"Supply Level 1σ: {High_Band_1std.round(2)}\n")
    print('-' * 165, '\n')
    print(f"Demand Level 1σ: {Low_Band_1std.round(2)}\n")
    print(f"Demand Level 2σ: {Low_Band_2std.round(2)}\n")
    print(f"Demand Level 3σ: {Low_Band_3std.round(2)}\n")
    
    
    # Creating the text output
    text_info = f"""
    Volatility-Based Supply and Demand Levels for {ticker} in {int(reference_year) + 1}\n
    Supply Level 3σ: {High_Band_3std.round(2)}\n
    Supply Level 2σ: {High_Band_2std.round(2)}\n
    Supply Level 1σ: {High_Band_1std.round(2)}\n
    -------------------------\n
    Demand Level 1σ: {Low_Band_1std.round(2)}\n
    Demand Level 2σ: {Low_Band_2std.round(2)}\n
    Demand Level 3σ: {Low_Band_3std.round(2)}\n
    """
    
    return fig, text_info

def wrapper_func(ticker, reference_year):
    try:
        fig, text_info = plot_volatility_bands(ticker, str(int(reference_year)))  
        return fig, text_info
    except Exception as e:
        error_message = str(e)
        return error_message, error_message

iface = gr.Interface(
    title = '📈 Volatility-Based Supply and Demand Levels 📉',
    description="""<br> The ideas presented in this app are based on the research paper <a href = "https://www.outspokenmarket.com/uploads/8/8/2/3/88233040/supply_and_demand_levels_forecasting_based_on_returns_volatility.pdf">Supply and Demand Levels Forecasting Based on Returns Volatility</a>,authored by <a href = "https://www.outspokenmarket.com/">Leandro Guerra</a>. <br>
                In the Kaggle Notebook, <a href = "https://www.kaggle.com/code/lusfernandotorres/volatility-based-supply-and-demand-levels/notebook">Volatility-Based Supply & Demand Levels Forecasting</a>, you can read further on how these supply and demand levels are calculated. <br><br>
                <b>Intructions</b><br>
                Enter the ticker of a security as it is displayed on Yahoo Finance. Then inform a year of reference. <br>
                The supply and demand levels are forecastings for the following year. If you inform <i>2022</i> as input, the plot will contain the forecastings for <i>2023</i>. <br><br>
                
                • Dot lines represent one standard deviation (68.7% probability). <br>

                • Dash lines represent two standard deviations (95.4% probability). <br>

                • Dash and dot lines represent three standard deviations (99.7% probability).""",
    fn=wrapper_func,
    inputs=[
        gr.Textbox(label="Enter the Ticker as it Appears on Yahoo Finance"),
        gr.Number(label="Enter the Year of Reference")
    ],
    outputs=[
        gr.Plot(label = ""),
        gr.Textbox(label = "")
    ],
    css=".gr-input {width: 60px;}"
)

iface.launch()