Portfolio_consruction_interactively_basic.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# 2 Portfolio consruction interactively"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Goal: User can change parameters in a box below. Then observe the performance of the stock portfolio. The strategy is 'healthy reversal' which aims to pick stocks experiencing price decline, but that are healthy otherwise. Parameters that can be changed: upper quantile of previous returns, minimum EBIT margin, and minimum revenue growth.  <br>\n",
    "Data - in this notebook is downloaded from Dropbox. The code for obtaining data is run in notebook 'Plot1_featured_predictive_model' "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    " ++++++++++++++++++++++++++++++++++++++++++++ <br>\n",
    " By: Yuriy Podvysotskiy <br>\n",
    " Date: November 4, 2019 <br>\n",
    " ++++++++++++++++++++++++++++++++++++++++++++"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import packages:\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Get the collected data:\n",
    "id1 = \"https://www.dropbox.com/s/x0tzza87nrie9rr/SP100_panel.csv?dl=1\"\n",
    "id2 = \"https://www.dropbox.com/s/nmyyrs7yehh7gdo/SP100.csv?dl=1\"\n",
    "\n",
    "df_panel = pd.read_csv(id1)\n",
    "df_mkt = pd.read_csv(id2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Function for making analysis (run it)\n",
    "def backtest_strategy(df_panel, df_mkt, ret_quant, low_ebitm, low_revg):\n",
    "    df_panel['ret_bottom'] = 0\n",
    "    df_panel['return'] = df_panel['return'].fillna(0)\n",
    "    \n",
    "    quantiles = df_panel.groupby(['date'])['return'].quantile(ret_quant)\n",
    "    quantiles = pd.DataFrame(quantiles)\n",
    "    quantiles.columns = ['low_quantile']\n",
    "    \n",
    "    quantiles['date'] = quantiles.index\n",
    "    quantiles = quantiles.reset_index(drop = True)\n",
    "    \n",
    "    df_panel = df_panel.merge(quantiles, how = 'left', on = ['date'])\n",
    "    df_panel['low_quantile'].describe()\n",
    "    \n",
    "    df_panel.loc[df_panel['return'] <= df_panel['low_quantile'],'ret_bottom'] = 1\n",
    "    \n",
    "    df_panel['perf_well'] = 0\n",
    "    df_panel.loc[(df_panel['ebitm'] > low_ebitm) & (df_panel['revg'] > low_revg), 'perf_well'] = 1\n",
    "    \n",
    "    df_panel['chosen'] = df_panel['ret_bottom'] * df_panel['perf_well']\n",
    "    df_panel[['ret_bottom', 'perf_well', 'chosen']].describe()\n",
    "    \n",
    "    df_panel = df_panel.sort_values(by = ['ticker', 'date'])\n",
    "    df_panel = df_panel.reset_index(drop = True)\n",
    "    \n",
    "    df_panel['chosen1'] = df_panel.groupby(['ticker'])['chosen'].shift(1)\n",
    "    \n",
    "    \n",
    "    # 2) ******************** test strategy *************************************\n",
    "    ## 2.1) pick all stocks that correspond to the rule\n",
    "    n_sums = df_panel.groupby(['date'])['chosen1'].sum()\n",
    "    n_sums = pd.DataFrame(n_sums)\n",
    "    n_sums.columns = ['n_sum']\n",
    "    \n",
    "    n_sums['date'] = n_sums.index\n",
    "    n_sums = n_sums.reset_index(drop = True)\n",
    "    \n",
    "    df_panel = df_panel.merge(n_sums, how = 'left', on = ['date'])\n",
    "  \n",
    "    df_panel['portf_ret'] = df_panel['return'] * df_panel['chosen1'] / df_panel['n_sum']\n",
    "    df_panel['portf_ret'] = df_panel['portf_ret'].fillna(0)\n",
    "    \n",
    "    df_panel = df_panel[['date', 'portf_ret']]\n",
    "    df_panel = df_panel.dropna()\n",
    "    \n",
    "    df_portf = df_panel.groupby(['date']).sum()\n",
    "    df_portf['date'] = df_portf.index\n",
    "    df_portf = df_portf.reset_index(drop = True)\n",
    "    df_portf = df_portf.merge(df_mkt, on = ['date'])\n",
    "    df_portf = df_portf.fillna(method = 'ffill')\n",
    "    df_portf = df_portf.sort_values(by = ['date'])\n",
    "    df_portf = df_portf.reset_index(drop = True)\n",
    "    \n",
    "    df_portf['portf_cumret'] = df_portf['portf_ret'].cumsum()\n",
    "    df_portf['mkt_cumret'] = df_portf['mkt_ret'].cumsum()\n",
    "    return(df_portf)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Comment: \n",
      "\n",
      " The user can change the three parameters above, \n",
      "  and observe the change in the dynamics of the portfolio peformance versus the relevant market index. \n",
      " Also, below is the quantitative evaluation based on the key investment performance metrics (Sharpe, Alpha, and Beta).\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.legend.Legend at 0x291e0400e48>"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# User parameters (please change!):\n",
    "ret_quant0 = 0.25  # pick stocks with returns below given quantile (for reversion strategy)\n",
    "low_ebitm0 = 0.1   # pick stocks with ebit margin above given threshold\n",
    "low_revg0 = 0.05   # pick stocks with revenue growth above certain threshold\n",
    "\n",
    "# ----------------------------------------------------------------------\n",
    "# Run the portfolio formation:\n",
    "df_portf = backtest_strategy(df_panel, df_mkt, \n",
    "                             ret_quant = ret_quant0, low_ebitm = low_ebitm0, low_revg = low_revg0)\n",
    "\n",
    "print(\"Comment: \\n\\n\", \"The user can change the three parameters above, \\n\", \n",
    "      \" and observe the change in the dynamics of the portfolio peformance versus the relevant market index. \\n\", \n",
    "     \"Also, below is the quantitative evaluation based on the key investment performance metrics (Sharpe, Alpha, and Beta).\")\n",
    "\n",
    "# And see the portfolio performance vs. the parket:\n",
    "ax = df_portf.plot(x = \"date\", y = ['portf_cumret', 'mkt_cumret'])\n",
    "ax.set_ylabel(\"Cumulative return\")\n",
    "ax.set_xlabel(\"\")\n",
    "ax.legend([\"Portfolio\", \"S&P 100\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Comment: \n",
      "\n",
      " The constructed portoflio has higher annualized return, but also higher risk. \n",
      " Its risk-adjusted return (Alpha) is positive 10.1% on annualized basis. \n",
      " Sharpe ratio is worse than that of the market. \n",
      "\n",
      " BUT the goal here is to design user-interaction system. Which can then be used for constructing superior portfolios.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Portfolio</th>\n",
       "      <th>Market</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>return</td>\n",
       "      <td>0.2415</td>\n",
       "      <td>0.1092</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>st dev</td>\n",
       "      <td>0.3200</td>\n",
       "      <td>0.1190</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>sharpe</td>\n",
       "      <td>0.6923</td>\n",
       "      <td>0.7499</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>alpha</td>\n",
       "      <td>0.1010</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>beta</td>\n",
       "      <td>1.2862</td>\n",
       "      <td>1.0000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       Portfolio  Market\n",
       "return    0.2415  0.1092\n",
       "st dev    0.3200  0.1190\n",
       "sharpe    0.6923  0.7499\n",
       "alpha     0.1010  0.0000\n",
       "beta      1.2862  1.0000"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Evaluation of portfolio versus the market:\n",
    "from sklearn.linear_model import LinearRegression\n",
    "\n",
    "## Indicators:  Ann_ret, Ann_sd, Sharpe, Alpha, Beta\n",
    "## For:         Portfolio, Market \n",
    "\n",
    "portf_Y = df_portf['portf_ret']\n",
    "portf_X = df_portf['mkt_ret']\n",
    "\n",
    "reg_p = LinearRegression().fit(portf_X.values.reshape(-1, 1), portf_Y.values.reshape(-1, 1))\n",
    "\n",
    "df_perf1 = pd.DataFrame( columns = ['Portfolio', 'Market'], \n",
    "                       index = ['return', 'st dev', 'sharpe', 'alpha', 'beta'])\n",
    "\n",
    "df_perf1.loc['return', 'Portfolio'] = \"{:.4f}\".format(df_portf['portf_ret'].mean()*51)\n",
    "df_perf1.loc['st dev', 'Portfolio'] = \"{:.4f}\".format(df_portf['portf_ret'].std()*(51**0.5))\n",
    "df_perf1.loc['sharpe', 'Portfolio'] = \"{:.4f}\".format((df_portf['portf_ret'].mean()*51 - .02)/(df_portf['portf_ret'].std()*(51**0.5)))\n",
    "df_perf1.loc['alpha', 'Portfolio'] = \"{:.4f}\".format(reg_p.intercept_[0]*51)\n",
    "df_perf1.loc['beta', 'Portfolio'] = \"{:.4f}\".format(reg_p.coef_[0][0])\n",
    "\n",
    "df_perf1.loc['return', 'Market'] = \"{:.4f}\".format(df_portf['mkt_ret'].mean()*51)\n",
    "df_perf1.loc['st dev', 'Market'] = \"{:.4f}\".format(df_portf['mkt_ret'].std()*(51**0.5))\n",
    "df_perf1.loc['sharpe', 'Market'] = \"{:.4f}\".format((df_portf['mkt_ret'].mean()*51 - .02)/ (df_portf['mkt_ret'].std()*(51**0.5)))\n",
    "df_perf1.loc['alpha', 'Market'] = \"{:.4f}\".format(0)\n",
    "df_perf1.loc['beta', 'Market'] = \"{:.4f}\".format(1)\n",
    "\n",
    "print(\"Comment: \\n\\n\", \"The constructed portoflio has higher annualized return, but also higher risk. \\n\",\n",
    "      \"Its risk-adjusted return (Alpha) is positive 10.1% on annualized basis. \\n\", \n",
    "     \"Sharpe ratio is worse than that of the market. \\n\\n\",\n",
    "     \"BUT\",\"the goal here is to design user-interaction system. Which can then be used for constructing superior portfolios.\")\n",
    "\n",
    "df_perf1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}