AFML 2.1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Structure\n",
    "\n",
    "This notebook will cover:\n",
    "\n",
    "Exercise 2.*\n",
    "\n",
    "Using different types of data structure to transform time-series into event-driven order.\n",
    "\n",
    "Instead of market price time-series, we form somewhat an transaction volume expectation to total transaction volume to resample data.\n",
    "\n",
    "Using such data structure could generate good market signal for different plausible quant strategies.\n",
    "\n",
    "In order to appreciate this technique, I highly recommend you to read the below research.\n",
    "\n",
    "[Volume Clock SSRN](https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2034858)\n",
    "\n",
    "**Note**\n",
    "\n",
    "If actual High-frequency data is not within your reach, you may wish to try with the below code:\n",
    "\n",
    "    def create_price_data(start_price: float = 1000.00,\n",
    "                          mu: float = .0,\n",
    "                          var: float = 1.0,\n",
    "                          n_samples: int = 1000000):\n",
    "                          \n",
    "        import numpy as np\n",
    "        import pandas as pd\n",
    "\n",
    "        i = np.random.normal(mu, var, n_samples)\n",
    "        df0 = pd.date_range(periods=n_samples,\n",
    "                            freq=pd.tseries.offsets.Minute(),\n",
    "                            end=pd.datetime.today())\n",
    "                            \n",
    "        X = pd.Series(i, index=df0, name = \"close\").to_frame()\n",
    "        X.close.iat[0] = start_price\n",
    "        X.cumsum().plot.line()\n",
    "        return X.cumsum()\n",
    "\n",
    "The above function can generate about 2 years of synthetic raw data, so that you can start to structure your data into dollar bars.\n",
    "\n",
    "Please remember to save this as your sample data to csv format, otherwise, your result may not be consistant. (This sample data can get you started for 1st 5 chapters)\n",
    "\n",
    "Contact: boyboi86@gmail.com"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "p = print\n",
    "\n",
    "#pls take note of version\n",
    "#numpy 1.17.3\n",
    "#pandas 1.0.3\n",
    "#sklearn 0.21.3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Intraday sample data courtesy of mlfinlab\n",
    "\n",
    "dollar = pd.read_csv('./Sample_data/dollar_bars.txt', \n",
    "                 sep=',', \n",
    "                 header=0, \n",
    "                 parse_dates = True, \n",
    "                 index_col=['date_time'])\n",
    "\n",
    "volume = pd.read_csv('./Sample_data/volume_bars.txt', \n",
    "                 sep=',', \n",
    "                 header=0, \n",
    "                 parse_dates = True, \n",
    "                 index_col=['date_time'])\n",
    "\n",
    "tick = pd.read_csv('./Sample_data/tick_bars.txt', \n",
    "                 sep=',', \n",
    "                 header=0, \n",
    "                 parse_dates = True, \n",
    "                 index_col=['date_time'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2042ce375c8>"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1800x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "db_ = dollar['close'].resample('W').count()\n",
    "vb_ = volume['close'].resample('W').count()\n",
    "tb_ = tick['close'].resample('W').count()\n",
    "\n",
    "count_df = pd.concat([tb_, vb_, db_], axis=1)\n",
    "count_df.columns = ['tick', 'volume', 'dollar']\n",
    "count_df.loc[:, ['tick', 'volume', 'dollar']].plot(kind = 'bar', figsize=[25, 5])\n",
    "\n",
    "# Tick bars have the most irregular count\n",
    "# While Dollar bar produces the most stable count per week"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.3082662585319557\n",
      "0.3653110836430761\n",
      "0.7012168586776739\n"
     ]
    }
   ],
   "source": [
    "p(count_df['dollar'].autocorr())\n",
    "p(count_df['volume'].autocorr())\n",
    "p(count_df['tick'].autocorr())\n",
    "\n",
    "#Dollar bars has the lowest autocorr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.0008405980471489872 0.0008367515769979635 0.0008319813710860616\n"
     ]
    }
   ],
   "source": [
    "db1_ = dollar['close'].resample('M').mean().pct_change().var()\n",
    "vb1_ = volume['close'].resample('M').mean().pct_change().var()\n",
    "tb1_ = tick['close'].resample('M').mean().pct_change().var()\n",
    "\n",
    "p(tb1_, vb1_, db1_)\n",
    "\n",
    "# Still dollar bar has the lowest variance\n",
    "# But i suspect you have to resample 1D and 1W as well"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "63093.465360938826 82208.37890653765 75737.39864288477\n"
     ]
    }
   ],
   "source": [
    "from scipy import stats\n",
    "\n",
    "p(stats.jarque_bera(dollar['close'].pct_change().dropna())[0],\n",
    "stats.jarque_bera(volume['close'].pct_change().dropna())[0],\n",
    "stats.jarque_bera(tick['close'].pct_change().dropna())[0])\n",
    "\n",
    "# Again.. dollar bar.. we r seeing a pattern here"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\Wei_X\\Anaconda3\\lib\\site-packages\\statsmodels\\tools\\_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n",
      "  import pandas.util.testing as tm\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total count: 1836\n",
      "Upper Bound exceed: 907\n",
      "Lower Bound exceed: 929\n"
     ]
    }
   ],
   "source": [
    "import statsmodels.stats.diagnostic as sm\n",
    "import statsmodels.api as smi\n",
    "\n",
    "def bband(data: pd.Series, window: int = 21, width: float = 0.005):\n",
    "    avg = data.ewm(span = window).mean()\n",
    "    std0 = avg * width\n",
    "    lower = avg - std0\n",
    "    upper = avg + std0\n",
    "    \n",
    "    return avg, upper, lower, std0\n",
    "\n",
    "dollar['avg'], dollar['upper'], dollar['lower'], dollar['std0'] = bband(dollar['close'])\n",
    "\n",
    "\n",
    "count_dn = dollar[dollar['lower'] > dollar['close']]\n",
    "count_up = dollar[dollar['upper'] < dollar['close']]\n",
    "bband_dollar = pd.concat([count_dn, count_up])\n",
    "\n",
    "raw_bbd = bband_dollar.copy()\n",
    "\n",
    "p(\"Total count: {0}\\nUpper Bound exceed: {1}\\nLower Bound exceed: {2}\".format(bband_dollar.count()[0], \n",
    "                                                                              count_up.count()[0], \n",
    "                                                                              count_dn.count()[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total count after filter (close price): 1782\n",
      "Total count after filter (absolute change): 424\n",
      "Total count after filter (pct change): 426\n"
     ]
    }
   ],
   "source": [
    "# when you import research as rs\n",
    "# the below func can be used as rs.cs_filters()\n",
    "\n",
    "def cumsum_events(df: pd.Series, limit: float):\n",
    "    idx, _up, _dn = [], 0, 0\n",
    "    diff = df.diff()\n",
    "    for i in diff.index[1:]:\n",
    "        _up, _dn = max(0, _up + diff.loc[i]), min(0, _dn + diff.loc[i])\n",
    "        if _up > limit:\n",
    "            _up = 0; idx.append(i)\n",
    "        elif _dn < - limit:\n",
    "            _dn = 0; idx.append(i)\n",
    "        \n",
    "    return pd.DatetimeIndex(idx)\n",
    "\n",
    "def cumsum_events1(df: pd.Series, limit: float):\n",
    "    idx, _up, _dn = [], 0, 0\n",
    "    diff = df.pct_change()\n",
    "    for i in diff.index[1:]:\n",
    "        _up, _dn = max(0, _up + diff.loc[i]), min(0, _dn + diff.loc[i])\n",
    "        if _up > limit:\n",
    "            _up = 0; idx.append(i)\n",
    "        elif _dn < - limit:\n",
    "            _dn = 0; idx.append(i)\n",
    "        \n",
    "    return pd.DatetimeIndex(idx)\n",
    "\n",
    "event = cumsum_events(bband_dollar['close'], limit = 0.005) # benchmark\n",
    "event_pct = cumsum_events1(bband_dollar['close'], limit = 0.005)\n",
    "event_abs = cumsum_events(bband_dollar['close'], limit = bband_dollar['std0'].mean()) # based on ewma std abs estimate 0.005\n",
    "\n",
    "event_count0 = dollar.reindex(event)\n",
    "event_count1 = dollar.reindex(event_abs)\n",
    "event_count2 = dollar.reindex(event_pct)\n",
    "\n",
    "p(\"Total count after filter (close price): {0}\".format(event_count0.count()[0]))\n",
    "p(\"Total count after filter (absolute change): {0}\".format(event_count1.count()[0]))\n",
    "p(\"Total count after filter (pct change): {0}\".format(event_count2.count()[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## White Test\n",
    "\n",
    "Heteroscedasticity tests imply the two following hypotheses.\n",
    "\n",
    "H0 (null hypothesis): data is homoscedastic.\n",
    "\n",
    "Ha (alternative hypothesis): data is heteroscedastic.\n",
    "\n",
    "Therefore, if the p-value associated to a heteroscedasticity test falls below a certain threshold (0.05 for example), we would conclude that the data is significantly heteroscedastic."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "White-Test p-Value: 1.3704431992973026e-08\n",
      "White test outcome at 5% signficance: heteroscedastic\n",
      "White-Test p-Value: 3.1055617824761596e-08\n",
      "White test outcome at 5% signficance: heteroscedastic\n"
     ]
    }
   ],
   "source": [
    "#event_count['std'] = event_count['close'].rolling(21).std()\n",
    "#event_count.dropna(inplace= True)\n",
    "\n",
    "def white_test(data: pd.DataFrame, window: int = 21):\n",
    "    data['std1'] = data['close'].rolling(21).std()\n",
    "    data.dropna(inplace= True)\n",
    "    X = smi.tools.tools.add_constant(data['close'])\n",
    "    results = smi.regression.linear_model.OLS(data['std1'], X).fit()\n",
    "    resid = results.resid\n",
    "    exog = results.model.exog\n",
    "    p(\"White-Test p-Value: {0}\".format(sm.het_white(resid, exog)[1]))\n",
    "    if sm.het_white(resid, exog)[1] > 0.05:\n",
    "        p(\"White test outcome at 5% signficance: homoscedastic\")\n",
    "    else:\n",
    "        p(\"White test outcome at 5% signficance: heteroscedastic\")\n",
    "\n",
    "# Without cumsum filter percentage return based on boillinger band would be more heteroscedastic\n",
    "# Main reason would be because it would filter out those signal that does not meet threshold requirement.\n",
    "\n",
    "white_test(raw_bbd) # without filter (less heteroscedastic)\n",
    "white_test(event_count0) # with filter (close price)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "White-Test p-Value: 0.27030043663183695\n",
      "White test outcome at 5% signficance: homoscedastic\n",
      "White-Test p-Value: 0.07936943626566823\n",
      "White test outcome at 5% signficance: homoscedastic\n"
     ]
    }
   ],
   "source": [
    "# As compared to percentage change vs absolute\n",
    "# Absolute change in daily price will yield lower p-value (more heteroscedastic)\n",
    "\n",
    "white_test(event_count1) # absolute return as a filter (less heteroscedastic)\n",
    "white_test(event_count2) # percent return as a filter"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Conclusion\n",
    "\n",
    "#### Statistical properties\n",
    "\n",
    "Dollar bars has the most stable frequency counts, closest to normality and low serial correlation.\n",
    "\n",
    "Other concerns are stationarity, volatility and moving average, which are not part of this exercise.\n",
    "\n",
    "#### White test/ Heteroscedasticity test\n",
    "\n",
    "Absolute return is preferred for cusum filter because the absolute change reflects actual dispersion within prices.\n",
    "\n",
    "Percentage return however causes short memory if any, plus it's relativity to previous value will cause clustering thus driving up heteroscedasticity.\n",
    "\n",
    "Absolute return should always be used and based on initial model indicator's std to achieve homoscedastic."
   ]
  }
 ],
 "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}