CoinbasePro_Portfolio_Optimizer.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "view-in-github"
   },
   "source": [
    "# Coinbase Pro Portfolio Optimization\n",
    "\n",
    "Adapted from original publication by [Brick Pop](https://medium.com/@brickpop):\n",
    "\n",
    "[Finding top crypto portfolios with Tensorflow and Matrix calculus](https://medium.com/stack-me-up/crypto-portfolio-optimization-with-python-and-tensorflow-an-approach-aa504578c799)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "B_1DpJleZwFG"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /opt/conda/lib/python3.9/site-packages/tensorflow/python/compat/v2_compat.py:96: disable_resource_variables (from tensorflow.python.ops.variable_scope) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "non-resource variables are not supported in the long term\n"
     ]
    }
   ],
   "source": [
    "import json\n",
    "import requests\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "#import tensorflow as tf\n",
    "import tensorflow.compat.v1 as tf\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "tf.disable_v2_behavior()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 2000
    },
    "colab_type": "code",
    "id": "NmozGK3Sq9aj",
    "outputId": "a0ff303c-b175-43cd-eed0-c34db5396a09"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Fetching: ETC\n",
      "Fetching: ETH\n",
      "Fetching: BTC\n",
      "Fetching: MKR\n",
      "Fetching: LTC\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>low</th>\n",
       "      <th>high</th>\n",
       "      <th>open</th>\n",
       "      <th>close</th>\n",
       "      <th>volume</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>time</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2021-06-14</th>\n",
       "      <td>38808.86</td>\n",
       "      <td>39249.00</td>\n",
       "      <td>39015.24</td>\n",
       "      <td>38871.34</td>\n",
       "      <td>288.033274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-13</th>\n",
       "      <td>34780.57</td>\n",
       "      <td>39396.00</td>\n",
       "      <td>35557.32</td>\n",
       "      <td>39015.24</td>\n",
       "      <td>15748.962040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-12</th>\n",
       "      <td>34635.47</td>\n",
       "      <td>37448.00</td>\n",
       "      <td>37340.08</td>\n",
       "      <td>35557.33</td>\n",
       "      <td>14516.545563</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-11</th>\n",
       "      <td>35944.00</td>\n",
       "      <td>37695.00</td>\n",
       "      <td>36694.91</td>\n",
       "      <td>37338.44</td>\n",
       "      <td>14142.193500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-10</th>\n",
       "      <td>35800.00</td>\n",
       "      <td>38425.67</td>\n",
       "      <td>37404.75</td>\n",
       "      <td>36694.05</td>\n",
       "      <td>19853.814189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-23</th>\n",
       "      <td>11525.62</td>\n",
       "      <td>11715.00</td>\n",
       "      <td>11672.92</td>\n",
       "      <td>11650.01</td>\n",
       "      <td>5516.170482</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-22</th>\n",
       "      <td>11370.00</td>\n",
       "      <td>11693.16</td>\n",
       "      <td>11529.99</td>\n",
       "      <td>11672.93</td>\n",
       "      <td>7946.024653</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-21</th>\n",
       "      <td>11480.00</td>\n",
       "      <td>11884.99</td>\n",
       "      <td>11864.26</td>\n",
       "      <td>11529.22</td>\n",
       "      <td>14865.378759</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-20</th>\n",
       "      <td>11671.48</td>\n",
       "      <td>11892.16</td>\n",
       "      <td>11758.87</td>\n",
       "      <td>11864.26</td>\n",
       "      <td>8801.177503</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-19</th>\n",
       "      <td>11569.00</td>\n",
       "      <td>12024.83</td>\n",
       "      <td>11954.22</td>\n",
       "      <td>11758.87</td>\n",
       "      <td>16646.314895</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>300 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 low      high      open     close        volume\n",
       "time                                                            \n",
       "2021-06-14  38808.86  39249.00  39015.24  38871.34    288.033274\n",
       "2021-06-13  34780.57  39396.00  35557.32  39015.24  15748.962040\n",
       "2021-06-12  34635.47  37448.00  37340.08  35557.33  14516.545563\n",
       "2021-06-11  35944.00  37695.00  36694.91  37338.44  14142.193500\n",
       "2021-06-10  35800.00  38425.67  37404.75  36694.05  19853.814189\n",
       "...              ...       ...       ...       ...           ...\n",
       "2020-08-23  11525.62  11715.00  11672.92  11650.01   5516.170482\n",
       "2020-08-22  11370.00  11693.16  11529.99  11672.93   7946.024653\n",
       "2020-08-21  11480.00  11884.99  11864.26  11529.22  14865.378759\n",
       "2020-08-20  11671.48  11892.16  11758.87  11864.26   8801.177503\n",
       "2020-08-19  11569.00  12024.83  11954.22  11758.87  16646.314895\n",
       "\n",
       "[300 rows x 5 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# GLOBAL\n",
    "# Coins currently in portfolio\n",
    "#coins = [\"BCH\", \"BTC\", \"DOGE\", \"ETH\", \"LTC\", \"USDT\"]\n",
    "coins = [\"ETC\", \"ETH\", \"BTC\", \"MKR\", \"LTC\"]\n",
    "# Adjust these two variables based on your currency and\n",
    "# time frame you are looking to trade in.\n",
    "primary_quote = 'BTC'\n",
    "quote_currecy = 'USD'\n",
    "granularity = 86400 # {60, 300, 900, 3600, 21600, 86400}\n",
    "\n",
    "# Not sure a start date is needed?\n",
    "start_date = '2020-08-17' # 2021-02-01\n",
    "\n",
    "days_ago_to_fetch = 400  # see also filter_history_by_date()\n",
    "coin_history = {}\n",
    "hist_length = 0\n",
    "average_returns = {}\n",
    "cumulative_returns = {}\n",
    "\n",
    "def fetch_all():\n",
    "  for coin in coins:\n",
    "    coin_history[coin] = fetch_history(coin)\n",
    "\n",
    "def fetch_history(coin):\n",
    "  print('Fetching: ' + coin)\n",
    "  #endpoint_url = \"https://min-api.cryptocompare.com/data/histoday?fsym={}&tsym=USD&limit={:d}\".format(coin, days_ago_to_fetch)\n",
    "  endpoint_url = f'https://api.pro.coinbase.com/products/{coin}-{quote_currecy}/candles?granularity={granularity}'\n",
    "  response = requests.get(endpoint_url)\n",
    "  hist = pd.DataFrame(json.loads(response.text), columns=['time', 'low', 'high', 'open', 'close', 'volume'])\n",
    "  hist = index_history(hist)\n",
    "  hist = filter_history_by_date(hist)\n",
    "  return hist\n",
    "\n",
    "def index_history(hist):\n",
    "  # index by date so we can easily filter by a given timeframe\n",
    "  hist = hist.set_index(['time'])\n",
    "  hist.index = pd.to_datetime(hist.index, unit='s')\n",
    "  return hist\n",
    "\n",
    "def filter_history_by_date(hist):\n",
    "  result = hist # customize here\n",
    "  # result = hist[hist.index.year >= 2017]\n",
    "  # result = result[result.index.day == 1] # every first of month, etc.\n",
    "  return result\n",
    "\n",
    "fetch_all()\n",
    "hist_length = len(coin_history[coins[0]])\n",
    "\n",
    "coin_history[primary_quote]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 2037
    },
    "colab_type": "code",
    "id": "72lz99gts8Nw",
    "outputId": "df23e56d-58d6-4dc2-fa22-e6b286162e45"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'ETC': 6.818914446640908, 'ETH': 4.9351985453016365, 'BTC': 2.2453979343838335, 'MKR': 3.514610007571341, 'LTC': 1.6133233728423448}\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>low</th>\n",
       "      <th>high</th>\n",
       "      <th>open</th>\n",
       "      <th>close</th>\n",
       "      <th>volume</th>\n",
       "      <th>return</th>\n",
       "      <th>excess_return</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>time</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2021-06-14</th>\n",
       "      <td>38808.86</td>\n",
       "      <td>39249.00</td>\n",
       "      <td>39015.24</td>\n",
       "      <td>38871.34</td>\n",
       "      <td>288.033274</td>\n",
       "      <td>-0.003688</td>\n",
       "      <td>-0.008491</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-13</th>\n",
       "      <td>34780.57</td>\n",
       "      <td>39396.00</td>\n",
       "      <td>35557.32</td>\n",
       "      <td>39015.24</td>\n",
       "      <td>15748.962040</td>\n",
       "      <td>0.097249</td>\n",
       "      <td>0.092447</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-12</th>\n",
       "      <td>34635.47</td>\n",
       "      <td>37448.00</td>\n",
       "      <td>37340.08</td>\n",
       "      <td>35557.33</td>\n",
       "      <td>14516.545563</td>\n",
       "      <td>-0.047744</td>\n",
       "      <td>-0.052546</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-11</th>\n",
       "      <td>35944.00</td>\n",
       "      <td>37695.00</td>\n",
       "      <td>36694.91</td>\n",
       "      <td>37338.44</td>\n",
       "      <td>14142.193500</td>\n",
       "      <td>0.017537</td>\n",
       "      <td>0.012735</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-10</th>\n",
       "      <td>35800.00</td>\n",
       "      <td>38425.67</td>\n",
       "      <td>37404.75</td>\n",
       "      <td>36694.05</td>\n",
       "      <td>19853.814189</td>\n",
       "      <td>-0.019000</td>\n",
       "      <td>-0.023802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-23</th>\n",
       "      <td>11525.62</td>\n",
       "      <td>11715.00</td>\n",
       "      <td>11672.92</td>\n",
       "      <td>11650.01</td>\n",
       "      <td>5516.170482</td>\n",
       "      <td>-0.001963</td>\n",
       "      <td>-0.006765</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-22</th>\n",
       "      <td>11370.00</td>\n",
       "      <td>11693.16</td>\n",
       "      <td>11529.99</td>\n",
       "      <td>11672.93</td>\n",
       "      <td>7946.024653</td>\n",
       "      <td>0.012397</td>\n",
       "      <td>0.007595</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-21</th>\n",
       "      <td>11480.00</td>\n",
       "      <td>11884.99</td>\n",
       "      <td>11864.26</td>\n",
       "      <td>11529.22</td>\n",
       "      <td>14865.378759</td>\n",
       "      <td>-0.028239</td>\n",
       "      <td>-0.033042</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-20</th>\n",
       "      <td>11671.48</td>\n",
       "      <td>11892.16</td>\n",
       "      <td>11758.87</td>\n",
       "      <td>11864.26</td>\n",
       "      <td>8801.177503</td>\n",
       "      <td>0.008963</td>\n",
       "      <td>0.004160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-19</th>\n",
       "      <td>11569.00</td>\n",
       "      <td>12024.83</td>\n",
       "      <td>11954.22</td>\n",
       "      <td>11758.87</td>\n",
       "      <td>16646.314895</td>\n",
       "      <td>-0.016342</td>\n",
       "      <td>-0.021144</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>300 rows × 7 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 low      high      open     close        volume    return  \\\n",
       "time                                                                         \n",
       "2021-06-14  38808.86  39249.00  39015.24  38871.34    288.033274 -0.003688   \n",
       "2021-06-13  34780.57  39396.00  35557.32  39015.24  15748.962040  0.097249   \n",
       "2021-06-12  34635.47  37448.00  37340.08  35557.33  14516.545563 -0.047744   \n",
       "2021-06-11  35944.00  37695.00  36694.91  37338.44  14142.193500  0.017537   \n",
       "2021-06-10  35800.00  38425.67  37404.75  36694.05  19853.814189 -0.019000   \n",
       "...              ...       ...       ...       ...           ...       ...   \n",
       "2020-08-23  11525.62  11715.00  11672.92  11650.01   5516.170482 -0.001963   \n",
       "2020-08-22  11370.00  11693.16  11529.99  11672.93   7946.024653  0.012397   \n",
       "2020-08-21  11480.00  11884.99  11864.26  11529.22  14865.378759 -0.028239   \n",
       "2020-08-20  11671.48  11892.16  11758.87  11864.26   8801.177503  0.008963   \n",
       "2020-08-19  11569.00  12024.83  11954.22  11758.87  16646.314895 -0.016342   \n",
       "\n",
       "            excess_return  \n",
       "time                       \n",
       "2021-06-14      -0.008491  \n",
       "2021-06-13       0.092447  \n",
       "2021-06-12      -0.052546  \n",
       "2021-06-11       0.012735  \n",
       "2021-06-10      -0.023802  \n",
       "...                   ...  \n",
       "2020-08-23      -0.006765  \n",
       "2020-08-22       0.007595  \n",
       "2020-08-21      -0.033042  \n",
       "2020-08-20       0.004160  \n",
       "2020-08-19      -0.021144  \n",
       "\n",
       "[300 rows x 7 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Calculate returns and excess returns\n",
    "\n",
    "def add_all_returns():\n",
    "  for coin in coins:\n",
    "    hist = coin_history[coin]\n",
    "    hist['return'] = (hist['close'] - hist['open']) / hist['open']\n",
    "    average = hist[\"return\"].mean()\n",
    "    average_returns[coin] = average\n",
    "    cumulative_returns[coin] = (hist[\"return\"] + 1).prod() - 1\n",
    "    hist['excess_return'] = hist['return'] - average\n",
    "    coin_history[coin] = hist\n",
    "\n",
    "add_all_returns()\n",
    "\n",
    "# display data\n",
    "print(cumulative_returns)\n",
    "coin_history[primary_quote]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop coins that don't have enough data\n",
    "new_coins = []\n",
    "new_coin_hist = {}\n",
    "\n",
    "for coin in coins:\n",
    "    #print(f'coin: {coin}\\tlen: {len(coin_history[coin])}')\n",
    "    if (len(coin_history[coin]) == hist_length):\n",
    "        new_coins.append(coin)\n",
    "        new_coin_hist[coin] = coin_history[coin]\n",
    "\n",
    "coins = new_coins\n",
    "coin_history = new_coin_hist\n",
    "\n",
    "#for coin in coins:\n",
    "    #print(f'{coin}\\t{len(coin_history[coin])}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 2000
    },
    "colab_type": "code",
    "id": "Sh7uTWigDj4r",
    "outputId": "eea86283-2c0e-4b5b-ea06-ceee100bb2fe"
   },
   "outputs": [
    {
     "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>ETC</th>\n",
       "      <th>ETH</th>\n",
       "      <th>BTC</th>\n",
       "      <th>MKR</th>\n",
       "      <th>LTC</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>time</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2021-06-14</th>\n",
       "      <td>-0.020285</td>\n",
       "      <td>-0.012260</td>\n",
       "      <td>-0.008491</td>\n",
       "      <td>-0.016597</td>\n",
       "      <td>-0.011568</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-13</th>\n",
       "      <td>0.058403</td>\n",
       "      <td>0.050672</td>\n",
       "      <td>0.092447</td>\n",
       "      <td>0.062227</td>\n",
       "      <td>0.051226</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-12</th>\n",
       "      <td>-0.037683</td>\n",
       "      <td>-0.000716</td>\n",
       "      <td>-0.052546</td>\n",
       "      <td>-0.004797</td>\n",
       "      <td>-0.011690</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-11</th>\n",
       "      <td>-0.028975</td>\n",
       "      <td>-0.054734</td>\n",
       "      <td>0.012735</td>\n",
       "      <td>-0.067737</td>\n",
       "      <td>-0.038371</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2021-06-10</th>\n",
       "      <td>-0.083106</td>\n",
       "      <td>-0.061209</td>\n",
       "      <td>-0.023802</td>\n",
       "      <td>-0.075019</td>\n",
       "      <td>-0.030324</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-23</th>\n",
       "      <td>-0.006488</td>\n",
       "      <td>-0.019352</td>\n",
       "      <td>-0.006765</td>\n",
       "      <td>-0.012528</td>\n",
       "      <td>0.000274</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-22</th>\n",
       "      <td>0.013670</td>\n",
       "      <td>0.011968</td>\n",
       "      <td>0.007595</td>\n",
       "      <td>0.054828</td>\n",
       "      <td>0.014773</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-21</th>\n",
       "      <td>-0.058932</td>\n",
       "      <td>-0.074679</td>\n",
       "      <td>-0.033042</td>\n",
       "      <td>-0.095467</td>\n",
       "      <td>-0.064742</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-20</th>\n",
       "      <td>0.013777</td>\n",
       "      <td>0.012741</td>\n",
       "      <td>0.004160</td>\n",
       "      <td>0.009749</td>\n",
       "      <td>0.005418</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2020-08-19</th>\n",
       "      <td>-0.073690</td>\n",
       "      <td>-0.042527</td>\n",
       "      <td>-0.021144</td>\n",
       "      <td>-0.044867</td>\n",
       "      <td>-0.056245</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>300 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 ETC       ETH       BTC       MKR       LTC\n",
       "time                                                        \n",
       "2021-06-14 -0.020285 -0.012260 -0.008491 -0.016597 -0.011568\n",
       "2021-06-13  0.058403  0.050672  0.092447  0.062227  0.051226\n",
       "2021-06-12 -0.037683 -0.000716 -0.052546 -0.004797 -0.011690\n",
       "2021-06-11 -0.028975 -0.054734  0.012735 -0.067737 -0.038371\n",
       "2021-06-10 -0.083106 -0.061209 -0.023802 -0.075019 -0.030324\n",
       "...              ...       ...       ...       ...       ...\n",
       "2020-08-23 -0.006488 -0.019352 -0.006765 -0.012528  0.000274\n",
       "2020-08-22  0.013670  0.011968  0.007595  0.054828  0.014773\n",
       "2020-08-21 -0.058932 -0.074679 -0.033042 -0.095467 -0.064742\n",
       "2020-08-20  0.013777  0.012741  0.004160  0.009749  0.005418\n",
       "2020-08-19 -0.073690 -0.042527 -0.021144 -0.044867 -0.056245\n",
       "\n",
       "[300 rows x 5 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Excess matrix\n",
    "\n",
    "excess_matrix = np.zeros((hist_length, len(coins)))\n",
    "\n",
    "for i in range(0, hist_length):\n",
    "  for idx, coin in enumerate(coins):\n",
    "    excess_matrix[i][idx] = coin_history[coin].iloc[i]['excess_return']\n",
    "\n",
    "# Display\n",
    "\n",
    "pretty_matrix = pd.DataFrame(excess_matrix, columns = coins)\n",
    "pretty_matrix.index = coin_history[coins[0]].index\n",
    "\n",
    "pretty_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 266
    },
    "colab_type": "code",
    "id": "lhAu2dfE16Ew",
    "outputId": "081a06b3-5bef-4cde-86f2-9a8ed0b01f1d"
   },
   "outputs": [
    {
     "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>ETC</th>\n",
       "      <th>ETH</th>\n",
       "      <th>BTC</th>\n",
       "      <th>MKR</th>\n",
       "      <th>LTC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.007868</td>\n",
       "      <td>0.002789</td>\n",
       "      <td>0.001618</td>\n",
       "      <td>0.002909</td>\n",
       "      <td>0.003631</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ETH</th>\n",
       "      <td>0.002789</td>\n",
       "      <td>0.003423</td>\n",
       "      <td>0.001769</td>\n",
       "      <td>0.003242</td>\n",
       "      <td>0.002976</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BTC</th>\n",
       "      <td>0.001618</td>\n",
       "      <td>0.001769</td>\n",
       "      <td>0.001754</td>\n",
       "      <td>0.001627</td>\n",
       "      <td>0.002058</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MKR</th>\n",
       "      <td>0.002909</td>\n",
       "      <td>0.003242</td>\n",
       "      <td>0.001627</td>\n",
       "      <td>0.006739</td>\n",
       "      <td>0.002781</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LTC</th>\n",
       "      <td>0.003631</td>\n",
       "      <td>0.002976</td>\n",
       "      <td>0.002058</td>\n",
       "      <td>0.002781</td>\n",
       "      <td>0.004223</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          ETC       ETH       BTC       MKR       LTC\n",
       "ETC  0.007868  0.002789  0.001618  0.002909  0.003631\n",
       "ETH  0.002789  0.003423  0.001769  0.003242  0.002976\n",
       "BTC  0.001618  0.001769  0.001754  0.001627  0.002058\n",
       "MKR  0.002909  0.003242  0.001627  0.006739  0.002781\n",
       "LTC  0.003631  0.002976  0.002058  0.002781  0.004223"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Variance co-variance matrix\n",
    "\n",
    "product_matrix = np.matmul(excess_matrix.transpose(), excess_matrix)\n",
    "var_covar_matrix = product_matrix / hist_length\n",
    "\n",
    "# Display\n",
    "\n",
    "pretty_matrix = pd.DataFrame(var_covar_matrix, columns=coins, index=coins)\n",
    "\n",
    "pretty_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 266
    },
    "colab_type": "code",
    "id": "nmHUAPCjSXAY",
    "outputId": "fdda6534-6df5-4af4-b548-7f8968ea51ff"
   },
   "outputs": [
    {
     "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>Std Dev</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.088701</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ETH</th>\n",
       "      <td>0.058509</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BTC</th>\n",
       "      <td>0.041877</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MKR</th>\n",
       "      <td>0.082091</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LTC</th>\n",
       "      <td>0.064985</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Std Dev\n",
       "ETC  0.088701\n",
       "ETH  0.058509\n",
       "BTC  0.041877\n",
       "MKR  0.082091\n",
       "LTC  0.064985"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Standard Deviation\n",
    "\n",
    "std_deviations = np.zeros((len(coins), 1))\n",
    "\n",
    "for idx, coin in enumerate(coins):\n",
    "  std_deviations[idx][0] = np.std(coin_history[coin]['return'])\n",
    "  \n",
    "# Display\n",
    "\n",
    "pretty_matrix = pd.DataFrame(std_deviations, columns=['Std Dev'], index=coins)\n",
    "\n",
    "pretty_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 266
    },
    "colab_type": "code",
    "id": "P-xteA0pryNp",
    "outputId": "edd57aff-1d97-4e64-f6e6-6d9c29af31d0"
   },
   "outputs": [
    {
     "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>ETC</th>\n",
       "      <th>ETH</th>\n",
       "      <th>BTC</th>\n",
       "      <th>MKR</th>\n",
       "      <th>LTC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.007868</td>\n",
       "      <td>0.005190</td>\n",
       "      <td>0.003715</td>\n",
       "      <td>0.007282</td>\n",
       "      <td>0.005764</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ETH</th>\n",
       "      <td>0.005190</td>\n",
       "      <td>0.003423</td>\n",
       "      <td>0.002450</td>\n",
       "      <td>0.004803</td>\n",
       "      <td>0.003802</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BTC</th>\n",
       "      <td>0.003715</td>\n",
       "      <td>0.002450</td>\n",
       "      <td>0.001754</td>\n",
       "      <td>0.003438</td>\n",
       "      <td>0.002721</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MKR</th>\n",
       "      <td>0.007282</td>\n",
       "      <td>0.004803</td>\n",
       "      <td>0.003438</td>\n",
       "      <td>0.006739</td>\n",
       "      <td>0.005335</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LTC</th>\n",
       "      <td>0.005764</td>\n",
       "      <td>0.003802</td>\n",
       "      <td>0.002721</td>\n",
       "      <td>0.005335</td>\n",
       "      <td>0.004223</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          ETC       ETH       BTC       MKR       LTC\n",
       "ETC  0.007868  0.005190  0.003715  0.007282  0.005764\n",
       "ETH  0.005190  0.003423  0.002450  0.004803  0.003802\n",
       "BTC  0.003715  0.002450  0.001754  0.003438  0.002721\n",
       "MKR  0.007282  0.004803  0.003438  0.006739  0.005335\n",
       "LTC  0.005764  0.003802  0.002721  0.005335  0.004223"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Std Deviation products matrix\n",
    "\n",
    "sdev_product_matrix = np.matmul(std_deviations, std_deviations.transpose())\n",
    "\n",
    "\n",
    "# Display\n",
    "\n",
    "pretty_matrix = pd.DataFrame(sdev_product_matrix, columns=coins, index=coins)\n",
    "\n",
    "pretty_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 266
    },
    "colab_type": "code",
    "id": "DPjF9DZDzmES",
    "outputId": "dd3e7f95-aaa1-433b-9b6e-636b0c31f2ee"
   },
   "outputs": [
    {
     "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>ETC</th>\n",
       "      <th>ETH</th>\n",
       "      <th>BTC</th>\n",
       "      <th>MKR</th>\n",
       "      <th>LTC</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.537469</td>\n",
       "      <td>0.435516</td>\n",
       "      <td>0.399439</td>\n",
       "      <td>0.629974</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ETH</th>\n",
       "      <td>0.537469</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.721801</td>\n",
       "      <td>0.674922</td>\n",
       "      <td>0.782646</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BTC</th>\n",
       "      <td>0.435516</td>\n",
       "      <td>0.721801</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.473387</td>\n",
       "      <td>0.756068</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MKR</th>\n",
       "      <td>0.399439</td>\n",
       "      <td>0.674922</td>\n",
       "      <td>0.473387</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.521214</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LTC</th>\n",
       "      <td>0.629974</td>\n",
       "      <td>0.782646</td>\n",
       "      <td>0.756068</td>\n",
       "      <td>0.521214</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          ETC       ETH       BTC       MKR       LTC\n",
       "ETC  1.000000  0.537469  0.435516  0.399439  0.629974\n",
       "ETH  0.537469  1.000000  0.721801  0.674922  0.782646\n",
       "BTC  0.435516  0.721801  1.000000  0.473387  0.756068\n",
       "MKR  0.399439  0.674922  0.473387  1.000000  0.521214\n",
       "LTC  0.629974  0.782646  0.756068  0.521214  1.000000"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Correlation matrix\n",
    "\n",
    "correlation_matrix = var_covar_matrix / sdev_product_matrix\n",
    "\n",
    "# Display\n",
    "\n",
    "pretty_matrix = pd.DataFrame(correlation_matrix, columns=coins, index=coins)\n",
    "\n",
    "pretty_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1116
    },
    "colab_type": "code",
    "id": "oa1uwHwC-RPr",
    "outputId": "5adcd2eb-88f4-4bc8-9abc-a321430e7ba6"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[round 0]\n",
      "Weights [[0.19931208]\n",
      " [0.19948086]\n",
      " [0.19967734]\n",
      " [0.19936757]\n",
      " [0.19942713]]\n",
      "Volatility: 5.45%\n",
      "\n",
      "[round 1000]\n",
      "Weights [[-0.00059522]\n",
      " [-0.00041544]\n",
      " [-0.00024796]\n",
      " [-0.00052255]\n",
      " [-0.00046091]]\n",
      "Volatility: 0.01%\n",
      "\n",
      "[round 2000]\n",
      "Weights [[-0.00059522]\n",
      " [-0.00041544]\n",
      " [-0.00024796]\n",
      " [-0.00052255]\n",
      " [-0.00046091]]\n",
      "Volatility: 0.01%\n",
      "\n",
      "[round 3000]\n",
      "Weights [[-0.00059522]\n",
      " [-0.00041544]\n",
      " [-0.00024796]\n",
      " [-0.00052255]\n",
      " [-0.00046091]]\n",
      "Volatility: 0.01%\n",
      "\n",
      "[round 4000]\n",
      "Weights [[-0.00059522]\n",
      " [-0.00041544]\n",
      " [-0.00024796]\n",
      " [-0.00052255]\n",
      " [-0.00046091]]\n",
      "Volatility: 0.01%\n",
      "\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>Weight %</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.012801</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ETH</th>\n",
       "      <td>0.008935</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BTC</th>\n",
       "      <td>0.005333</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MKR</th>\n",
       "      <td>0.011238</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LTC</th>\n",
       "      <td>0.009912</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Weight %\n",
       "ETC  0.012801\n",
       "ETH  0.008935\n",
       "BTC  0.005333\n",
       "MKR  0.011238\n",
       "LTC  0.009912"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Optimize weights to minimize variance\n",
    "\n",
    "def minimize_volatility():\n",
    "\n",
    "  # Define the model\n",
    "  # Portfolio Volatility = Sqrt (Transpose (Wt.SD) * Correlation Matrix * Wt. SD)\n",
    "\n",
    "  coin_weights = tf.Variable(np.full((len(coins), 1), 1.0 / len(coins))) # our variables\n",
    "  weighted_std_devs = tf.multiply(coin_weights, std_deviations)\n",
    "\n",
    "  product_1 = tf.transpose(weighted_std_devs)\n",
    "  product_2 = tf.matmul(product_1, correlation_matrix)\n",
    "  \n",
    "  portfolio_variance = tf.matmul(product_2, weighted_std_devs)\n",
    "  portfolio_volatility = tf.sqrt(tf.reduce_sum(portfolio_variance))\n",
    "\n",
    "\n",
    "  # Run\n",
    "  learn_rate = 0.01\n",
    "  steps = 5000\n",
    "  \n",
    "  init = tf.global_variables_initializer()\n",
    "\n",
    "  # Training using Gradient Descent to minimize variance\n",
    "  train_step = tf.train.GradientDescentOptimizer(learn_rate).minimize(portfolio_volatility)\n",
    "\n",
    "  with tf.Session() as sess:\n",
    "    sess.run(init)\n",
    "    for i in range(steps):\n",
    "      sess.run(train_step)\n",
    "      if i % 1000 == 0 :\n",
    "        print(\"[round {:d}]\".format(i))\n",
    "        print(\"Weights\", coin_weights.eval())\n",
    "        print(\"Volatility: {:.2f}%\".format(portfolio_volatility.eval() * 100))\n",
    "        print(\"\")\n",
    "        \n",
    "    return coin_weights.eval()\n",
    "\n",
    "weights = minimize_volatility()\n",
    "\n",
    "pretty_weights = pd.DataFrame(weights * 100, index = coins, columns = [\"Weight %\"])\n",
    "pretty_weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 606
    },
    "colab_type": "code",
    "id": "4iKXZjRyvSde",
    "outputId": "f9c38bff-0419-4e70-b264-efc13148ec11"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[round 0]\n",
      "Weights [[0.1998587 ]\n",
      " [0.20002794]\n",
      " [0.20022496]\n",
      " [0.19991434]\n",
      " [0.19997406]]\n",
      "Volatility: 5.47%\n",
      "\n",
      "[round 2500]\n",
      "Weights [[0.        ]\n",
      " [0.        ]\n",
      " [1.00169213]\n",
      " [0.        ]\n",
      " [0.        ]]\n",
      "Volatility: 4.19%\n",
      "\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>Weight %</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ETH</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BTC</th>\n",
       "      <td>100.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MKR</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LTC</th>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "     Weight %\n",
       "ETC       0.0\n",
       "ETH       0.0\n",
       "BTC     100.0\n",
       "MKR       0.0\n",
       "LTC       0.0"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Optimize weights to minimize variance\n",
    "\n",
    "def minimize_volatility():\n",
    "\n",
    "  # Define the model\n",
    "  # Portfolio Volatility = Sqrt (Transpose (Wt.SD) * Correlation Matrix * Wt. SD)\n",
    "\n",
    "  coin_weights = tf.Variable(np.full((len(coins), 1), 1.0 / len(coins))) # our variables\n",
    "  weighted_std_devs = tf.multiply(coin_weights, std_deviations)\n",
    "\n",
    "  product_1 = tf.transpose(weighted_std_devs)\n",
    "  product_2 = tf.matmul(product_1, correlation_matrix)\n",
    "  \n",
    "  portfolio_variance = tf.matmul(product_2, weighted_std_devs)\n",
    "  portfolio_volatility = tf.sqrt(tf.reduce_sum(portfolio_variance))\n",
    "\n",
    "  # Constraints: sum([0..1, 0..1, ...]) = 1\n",
    "\n",
    "  lower_than_zero = tf.greater( np.float64(0), coin_weights )\n",
    "  zero_minimum_op = coin_weights.assign( tf.where (lower_than_zero, tf.zeros_like(coin_weights), coin_weights) )\n",
    "\n",
    "  greater_than_one = tf.greater( coin_weights, np.float64(1) )\n",
    "  unity_max_op = coin_weights.assign( tf.where (greater_than_one, tf.ones_like(coin_weights), coin_weights) )\n",
    "\n",
    "  result_sum = tf.reduce_sum(coin_weights)\n",
    "  unity_sum_op = coin_weights.assign(tf.divide(coin_weights, result_sum))\n",
    "  \n",
    "  constraints_op = tf.group(zero_minimum_op, unity_max_op, unity_sum_op)\n",
    "\n",
    "  # Run\n",
    "  learning_rate = 0.01\n",
    "  steps = 5000\n",
    "  \n",
    "  init = tf.global_variables_initializer()\n",
    "\n",
    "  # Training using Gradient Descent to minimize variance\n",
    "  optimize_op = tf.train.GradientDescentOptimizer(learning_rate).minimize(portfolio_volatility)\n",
    "\n",
    "  with tf.Session() as sess:\n",
    "    sess.run(init)\n",
    "    for i in range(steps):\n",
    "      sess.run(optimize_op)\n",
    "      sess.run(constraints_op)\n",
    "      if i % 2500 == 0 :\n",
    "        print(\"[round {:d}]\".format(i))\n",
    "        print(\"Weights\", coin_weights.eval())\n",
    "        print(\"Volatility: {:.2f}%\".format(portfolio_volatility.eval() * 100))\n",
    "        print(\"\")\n",
    "        \n",
    "    sess.run(constraints_op)\n",
    "    return coin_weights.eval()\n",
    "\n",
    "weights = minimize_volatility()\n",
    "\n",
    "pretty_weights = pd.DataFrame(weights * 100, index = coins, columns = [\"Weight %\"])\n",
    "pretty_weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 759
    },
    "colab_type": "code",
    "id": "1LBAekr4orKs",
    "outputId": "6344addf-9b53-48dd-e307-20d0b722f39d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[round 0]\n",
      "Volatility 0.06 %\n",
      "Return 325.58 %\n",
      "Sharpe ratio 53.89852036378944\n",
      "\n",
      "[round 2000]\n",
      "Volatility 0.06 %\n",
      "Return 549.48 %\n",
      "Sharpe ratio 91.04997132599031\n",
      "\n",
      "[round 4000]\n",
      "Volatility 0.06 %\n",
      "Return 546.77 %\n",
      "Sharpe ratio 90.72360076845548\n",
      "\n",
      "[round 6000]\n",
      "Volatility 0.06 %\n",
      "Return 546.77 %\n",
      "Sharpe ratio 90.72591726639926\n",
      "\n",
      "[round 8000]\n",
      "Volatility 0.06 %\n",
      "Return 548.14 %\n",
      "Sharpe ratio 90.87728075644068\n",
      "\n",
      "Volatility 0.06 %\n",
      "Return 523.72 %\n",
      "Sharpe ratio 87.76484015771433\n",
      "Took 22.816048s to complete\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>Weight %</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>ETC</th>\n",
       "      <td>31.173236</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ETH</th>\n",
       "      <td>58.617893</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>BTC</th>\n",
       "      <td>0.756887</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>MKR</th>\n",
       "      <td>2.583342</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>LTC</th>\n",
       "      <td>6.868642</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Weight %\n",
       "ETC  31.173236\n",
       "ETH  58.617893\n",
       "BTC   0.756887\n",
       "MKR   2.583342\n",
       "LTC   6.868642"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Optimize weights to maximize return/risk\n",
    "\n",
    "import time\n",
    "start = time.time()\n",
    "\n",
    "def maximize_sharpe_ratio():\n",
    "  \n",
    "  # Define the model\n",
    "  \n",
    "  # 1) Variance\n",
    "  \n",
    "  coin_weights = tf.Variable(tf.random_uniform((len(coins), 1), dtype=tf.float64)) # our variables\n",
    "  weighted_std_devs = tf.multiply(coin_weights, std_deviations)\n",
    "  \n",
    "  product_1 = tf.transpose(weighted_std_devs)\n",
    "  product_2 = tf.matmul(product_1, correlation_matrix)\n",
    "  \n",
    "  portfolio_variance = tf.matmul(product_2, weighted_std_devs)\n",
    "  portfolio_volatility = tf.sqrt(tf.reduce_sum(portfolio_variance))\n",
    "\n",
    "  \n",
    "  # 2) Return\n",
    "  \n",
    "  returns = np.full((len(coins), 1), 0.0) # same as coin_weights\n",
    "  for coin_idx in range(0, len(coins)):\n",
    "    returns[coin_idx] = cumulative_returns[coins[coin_idx]]\n",
    "  \n",
    "  portfolio_return = tf.reduce_sum(tf.multiply(coin_weights, returns))\n",
    "  \n",
    "  # 3) Return / Risk\n",
    "  \n",
    "  sharpe_ratio = tf.divide(portfolio_return, portfolio_volatility)\n",
    "  \n",
    "  # Constraints\n",
    "  \n",
    "  # all values positive, with unity sum\n",
    "  weights_sum = tf.reduce_sum(coin_weights)\n",
    "  constraints_op = coin_weights.assign(tf.divide(tf.abs(coin_weights), tf.abs(weights_sum) ))\n",
    "  \n",
    "  # Run\n",
    "  learning_rate = 0.0001\n",
    "  learning_rate = 0.0015\n",
    "  steps = 10000\n",
    "  \n",
    "  # Training using Gradient Descent to minimize cost\n",
    "  \n",
    "  optimize_op = tf.train.GradientDescentOptimizer(learning_rate, use_locking=True).minimize(tf.negative(sharpe_ratio))\n",
    "  #2# optimize_op = tf.train.AdamOptimizer(learning_rate, use_locking=True).minimize(tf.negative(sharpe_ratio))\n",
    "  #3# optimize_op = tf.train.AdamOptimizer(learning_rate=0.00005, beta1=0.9, beta2=0.999, epsilon=1e-08, use_locking=False).minimize(tf.negative(sharpe_ratio))\n",
    "  #4# optimize_op = tf.train.AdagradOptimizer(learning_rate=0.01, initial_accumulator_value=0.1, use_locking=False).minimize(tf.negative(sharpe_ratio))\n",
    "  \n",
    "  \n",
    "  init = tf.global_variables_initializer()\n",
    "  \n",
    "  with tf.Session() as sess:\n",
    "    ratios = np.zeros(steps)\n",
    "    returns = np.zeros(steps)\n",
    "    sess.run(init)\n",
    "    for i in range(steps):\n",
    "      sess.run(optimize_op)\n",
    "      sess.run(constraints_op)\n",
    "      ratios[i] = sess.run(sharpe_ratio)\n",
    "      returns[i] = sess.run(portfolio_return) * 100\n",
    "      if i % 2000 == 0 : \n",
    "        sess.run(constraints_op)\n",
    "        print(\"[round {:d}]\".format(i))\n",
    "        #print(\"Coin weights\", sess.run(coin_weights))\n",
    "        print(\"Volatility {:.2f} %\".format(sess.run(portfolio_volatility)))\n",
    "        print(\"Return {:.2f} %\".format(sess.run(portfolio_return)*100))\n",
    "        print(\"Sharpe ratio\", sess.run(sharpe_ratio))\n",
    "        print(\"\")\n",
    "    \n",
    "    sess.run(constraints_op)\n",
    "    # print(\"Coin weights\", sess.run(coin_weights))\n",
    "    print(\"Volatility {:.2f} %\".format(sess.run(portfolio_volatility)))\n",
    "    print(\"Return {:.2f} %\".format(sess.run(portfolio_return)*100))\n",
    "    print(\"Sharpe ratio\", sess.run(sharpe_ratio))\n",
    "    return sess.run(coin_weights)\n",
    "\n",
    "weights = maximize_sharpe_ratio()\n",
    "\n",
    "print(\"Took {:f}s to complete\".format(time.time() - start))\n",
    "pretty_weights = pd.DataFrame(weights * 100, index = coins, columns = [\"Weight %\"])\n",
    "\n",
    "pretty_weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "include_colab_link": true,
   "name": "Crypto portfolio optimization with Python and Tensorflow part 1.ipynb",
   "provenance": [],
   "version": "0.3.2"
  },
  "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": "3.9.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}