nb_010_playing-with-timeseries.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# nb 010: Playing with Timeseries\n",
    "\n",
    "I need to map an irregular series of timestamps to a series of values and interpolate it with piecewise interpolation. I.e. with pairs `(9:01AM, \"O\"), (9:10AM, \"V\")`, the value `9:03AM` should map to `O`.\n",
    "\n",
    "This can be done by keeping a sorted list of timestamps and indexing it by a strictly increasing time-key (for $O(1)$ indexing) or indexing by an arbitrary time-key (for $O(log(n))$ indexing.)\n",
    "\n",
    "But Pandas might implement this already, which would probably be much faster than any solution I came up with. I need to find out how to do that. Once I do that, I need to make timelines, and then extract charging events from these.\n",
    "\n",
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "index = [pd.Timestamp(\"9:00\"),\n",
    "         pd.Timestamp(\"9:10\"),\n",
    "         pd.Timestamp(\"9:12\"),\n",
    "         pd.Timestamp(\"9:30\"),\n",
    "         pd.Timestamp(\"9:31\"),\n",
    "         pd.Timestamp(\"9:52\")]\n",
    "\n",
    "data = [3, 4, 10, 7, 2, 12]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "ss = pd.Series(data=data, index=index)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2020-02-26 09:00:00     3\n",
       "2020-02-26 09:10:00     4\n",
       "2020-02-26 09:12:00    10\n",
       "2020-02-26 09:30:00     7\n",
       "2020-02-26 09:31:00     2\n",
       "2020-02-26 09:52:00    12\n",
       "dtype: int64"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ss[\"9:00\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "2020-02-26 09:00:00     3.0\n",
       "2020-02-26 09:05:00     NaN\n",
       "2020-02-26 09:10:00     4.0\n",
       "2020-02-26 09:12:00    10.0\n",
       "2020-02-26 09:30:00     7.0\n",
       "2020-02-26 09:31:00     2.0\n",
       "2020-02-26 09:52:00    12.0\n",
       "dtype: float64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Hacky Method 1 for indexing between samples\n",
    "\n",
    "tk = pd.Timestamp(\"9:05\")\n",
    "ss[tk] = float(\"nan\")\n",
    "ss = ss.sort_index()\n",
    "ss"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3.0"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ss = ss.interpolate(\"pad\")\n",
    "ss[tk] #should be 3, not 12."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "assert ss[tk] == ss[\"9:00\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "3.0"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Better method!\n",
    "idx = ss.index.get_loc(pd.Timestamp(\"9:02\"), method=\"ffill\")\n",
    "ss[idx]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "tk = pd.Timestamp(\"9:51\")\n",
    "idx = ss.index.get_loc(pd.Timestamp(tk), method=\"ffill\")\n",
    "val = ss[idx]\n",
    "\n",
    "assert val == 2.0\n",
    "\n",
    "# Awesome!! This works great."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}