timeline_car_scooter_brex_san_jose.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook demonstrates the result of the first round of data collection, for the Mountain View library <-> San Jose library round trip."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Import all the dependencies"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# for reading and validating data\n",
    "import emeval.input.spec_details as eisd\n",
    "import emeval.input.phone_view as eipv\n",
    "import emeval.input.eval_view as eiev"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# Visualization helpers\n",
    "import emeval.viz.phone_view as ezpv\n",
    "import emeval.viz.eval_view as ezev"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# For plots\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# For maps\n",
    "import branca.element as bre"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Load and validate data\n",
    "\n",
    "The first issue to note is that we actually have two specs here. The first spec is the checked in `evaluation.spec.sample`, which defines calibration for both stationary and moving instances, and some evaluation trips. However, while starting with the calibration, we noticed some inconsistencies between the power curves. So in order to be more consistent, I defined a second, calibration-only spec `examples/calibration.only.json`, which essentially repeats the calibration experiments multiple times.\n",
    "\n",
    "After that, I returned to the first set of experiments for the moving calibration and the evaluation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "DATASTORE_URL = \"http://cardshark.cs.berkeley.edu\"\n",
    "AUTHOR_EMAIL = \"shankari@eecs.berkeley.edu\"\n",
    "sdunp = eisd.SpecDetails(DATASTORE_URL, AUTHOR_EMAIL, \"car_scooter_brex_san_jose\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import importlib\n",
    "importlib.reload(ezev)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pvunp = eipv.PhoneView(sdunp)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import arrow\n",
    "for t in pvunp.map()[\"ios\"][\"ucb-sdb-ios-1\"][\"transitions\"]:\n",
    "    if t[\"data\"][\"transition\"] in [\"START_EVALUATION_PERIOD\", \"STOP_EVALUATION_PERIOD\"] and t[\"data\"][\"spec_id\"] == sdunp.CURR_SPEC_ID:\n",
    "        print(t[\"_id\"], t[\"data\"][\"transition\"], t[\"data\"][\"trip_id\"], arrow.get(t[\"data\"][\"ts\"]).to(sdunp.eval_tz))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "pvunp.validate()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We see these evaluation trips have labels such as `HAHFDC_0`, or `HAMFDC_0`. What do these mean?\n",
    "\n",
    "`HA` or `MA` refer to high accuracy or medium accuracy, respectively. Trips with `HA` will tend to favor GPS utilization and result in high power consumption.\n",
    "\n",
    "`HF` or `MF` refer to high frequency or medium accuracy, respectively. Trips with `HF` will sense and process more often, and are likely to have higher spatiotemporal accuracy (e.g. will hug corners more accurately), albeit with higher power consumption."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "evunp = eiev.EvaluationView()\n",
    "evunp.from_view_eval_trips(pvunp, \"\", \"\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Display trips\n",
    "\n",
    "Click on the layer to toggle the lines"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "map_list = ezev.get_map_list_eval_trips(evunp, \"android\", \"\", \"HAHFDC|HAMFDC|MAHFDC|MAMFDC\")\n",
    "rows = ezpv.get_row_count(len(map_list), 2)\n",
    "evaluation_maps = bre.Figure(ratio=\"{}%\".format((rows/4) * 100))\n",
    "for i, curr_map in enumerate(map_list):\n",
    "    evaluation_maps.add_subplot(rows, 2, i+1).add_child(curr_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "evaluation_maps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "map_list = ezev.get_map_list_eval_trips(evunp, \"ios\", \"\", \"HAHFDC|HAMFDC|MAHFDC|MAMFDC\")\n",
    "rows = ezpv.get_row_count(len(map_list), 2)\n",
    "evaluation_maps = bre.Figure(ratio=\"{}%\".format((rows/4) * 100))\n",
    "for i, curr_map in enumerate(map_list):\n",
    "    evaluation_maps.add_subplot(rows, 2, i+1).add_child(curr_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "evaluation_maps"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Display sections"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "map_list = ezev.get_map_list_eval_sections(evunp, \"android\", \"\", \"HAHFDC|HAMFDC|MAHFDC|MAMFDC\")\n",
    "rows = ezpv.get_row_count(len(map_list), 2)\n",
    "evaluation_maps = bre.Figure(ratio=\"{}%\".format((rows/4) * 100))\n",
    "for i, curr_map in enumerate(map_list):\n",
    "    evaluation_maps.add_subplot(rows, 2, i+1).add_child(curr_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "evaluation_maps"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "map_list = ezev.get_map_list_eval_sections(evunp, \"ios\", \"\", \"HAHFDC|HAMFDC|MAHFDC|MAMFDC\")\n",
    "rows = ezpv.get_row_count(len(map_list), 2)\n",
    "evaluation_maps = bre.Figure(ratio=\"{}%\".format((rows/4) * 100))\n",
    "for i, curr_map in enumerate(map_list):\n",
    "    evaluation_maps.add_subplot(rows, 2, i+1).add_child(curr_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "evaluation_maps"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Display battery"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "(ifig, [android_ax, ios_ax]) = plt.subplots(ncols=1, nrows=2, figsize=(10,10))\n",
    "\n",
    "ezpv.plot_all_power_drain(ios_ax, pvunp.map()[\"ios\"], \"evaluation\", \"\")\n",
    "ios_ax.legend(loc=\"lower left\", bbox_to_anchor=(0, -0.75, 0, -0.5), ncol=2)\n",
    "ezpv.plot_all_power_drain(android_ax, pvunp.map()[\"android\"], \"evaluation\", \"\")\n",
    "android_ax.legend(loc=\"upper left\", bbox_to_anchor=(0,1.75), ncol=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "(ifig, [android_ax, ios_ax]) = plt.subplots(ncols=1, nrows=2, figsize=(10,10))\n",
    "\n",
    "color_map = {}\n",
    "legend_map = {}\n",
    "\n",
    "(color_map, legend_map) = ezpv.plot_collapsed_all_power_drain(ios_ax, pvunp.map()[\"ios\"], \"evaluation\", \"\", color_map, legend_map)\n",
    "(color_map, legend_map) = ezpv.plot_collapsed_all_power_drain(android_ax, pvunp.map()[\"android\"], \"evaluation\", \"\", color_map, legend_map)\n",
    "ios_ax.legend(legend_map.values(), legend_map.keys())\n",
    "android_ax.legend(legend_map.values(), legend_map.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "android_density_df = ezpv.filter_density_df(ezpv.get_location_density_df(pvunp.map()[\"android\"], \"evaluation\"))\n",
    "android_ax = android_density_df.plot(kind='density', subplots=False, figsize=(8,8), sharex=True, sharey=True)\n",
    "android_ax.legend(loc=\"center left\", bbox_to_anchor=(1, 0.5))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ios_density_df = ezpv.filter_density_df(ezpv.get_location_density_df(pvunp.map()[\"ios\"], \"evaluation\"))\n",
    "ios_ax = ios_density_df.plot(kind='density', subplots=False, figsize=(10,10), sharex=True, sharey=True)\n",
    "ios_ax.legend(loc=\"center left\", bbox_to_anchor=(1, 0.5))"
   ]
  }
 ],
 "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": "3.8.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}