conference travel co2.ipynb

conference-travel-co2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/gist/jacobeisenstein/ae0e13e270f3b00c9c2046b52297d018/conference-travel-co2.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# CO2 Emissions for ACL 2023\n",
        "\n",
        "* Jacob Eisenstein & Roy Schwartz, December 2023\n",
        "\n",
        "The carbon footprint of AI training and inference is deservedly receiving increasing attention. However, large models are only one way in which AI contributes to global warming: more prosaically, AI's [exponentially growing conference scene](https://thegradient.pub/neurips-2019-too-big/) involves thousands of flights all over the world.\n",
        "\n",
        "This notebook contains code to estimate the CO2 emissions associated with air travel to ACL 2023. We estimate that ACL's **2500 attendees** flew **23.6 million kilometers**, amounting to approximately **2400 metric tons** of CO2 emissions. To put this in perspective, it is equivalent to the emissions of:\n",
        "* Powering **1572 A100 GPUs for an entire year** at maximum load (see calculations below)\n",
        "* **512 person-years** for the average human being at the current global per-capita emissions rate.\n",
        "* Training [Llama](https://arxiv.org/pdf/2302.13971.pdf) 2.4 times; training [Llama 2](https://arxiv.org/abs/2307.09288) 4.4 times. (This is based on reported emissions of 1000 mT co2e and 540 mT co2e respectively.)\n",
        "\n"
      ],
      "metadata": {
        "id": "uz5iPtVqFFPj"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Methodology\n",
        "\n",
        "Our estimates are based on per-attendee city-level addresses (provided by the ACL), in combination with publicly-available flight data.\n",
        "\n",
        "* **Airports**: For each participant, we identify a set of candidate airports near their self-reported city-level home address.\n",
        "* **Flights**: We assume participants take a flight with the fewest connections, and that among flights with equal numbers of connections, take the one with the shortest combined distance.\n",
        "* **Emissions**: Emissions per flight can be calculated in two ways: using a piecewise linear function from [ICCT](https://theicct.org/sites/default/files/publications/ICCT_CO2-commercl-aviation-2018_20190918.pdf), or using a quadratic formula from [myclimate](https://www.myclimate.org/en/information/about-myclimate/downloads/flight-emission-calculator/).\n",
        "\n",
        "More details on each step are given in the code below."
      ],
      "metadata": {
        "id": "Ql3Jd3t9MYmt"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "## Previous estimates\n",
        "\n",
        "There have been several prior estimates for emissions associated with conferences, although none this thorough for ACL.\n",
        "\n",
        "- [Caines and Rei (2019)](https://www.marekrei.com/blog/geographic-diversity-of-nlp-conferences/) estimated **2100 metric tons** (mT) of CO2E from **10.5m km** of air travel for all NLP conferences in 2019. Their estimate was less precise in a few ways: for example, it was based on authors rather than attendees, and did not incorporate information about flight routes. Their estimate was based on 200g co2e per passenger-km, which was higher than the estimates in the sources we found.\n",
        "- [Skiles et al (2021)](https://www.nature.com/articles/s41893-021-00823-2#Sec22) estimated **3900 tons** co2e for ICLR 2019 (in New Orleans), roughly **1.5 tons per attendee**. However, this was based on only country-level geolocation per attendee, using the geographical center of the country. They had per-city attendee origins for two other conferences (NAMS and AAS), and estimated per-attendee emissions of roughly 1 ton co2e. Emissions were computed using [myclimate](https://www.myclimate.org/en/information/about-myclimate/downloads/flight-emission-calculator/), using the great circle distance between attendee origin and the conference location.\n",
        "- [Yakar and Kwee (2020)](https://www.sciencedirect.com/science/article/pii/S0720048X20300589#bib0040) estimated 39.5 mT co2e for 22000 attendees to a radiology conference in Chicago in 2017, working out to roughly **1.8 mT co2e per participant**. This estimate was based on state- and country-level geographical locations per participant. They also used the myclimate calculator.\n",
        "- [Jäckle (2022)](https://link.springer.com/chapter/10.1007/978-981-16-4911-0_2) examined a series of political science conferences, estimating emissions between **1-4 mT co2e per participant** depending on the location and the estimated greenhouse gas intensity of air travel. The estimate was based on the great circle route distance between the conference location and the home institution of paper presenters at the conference.\n",
        "- [Przybyła and Shardlow (2022)](https://aclanthology.org/2022.findings-acl.304.pdf) use the addresses of ACL anthology authors to track emissions over time. They estimate roughly **1 tonne** co2e per participant for most conferences, with total emissions rising over time with the number of papers. The emissions per flight were based on a regression to account for change in efficiency over time. Similar to other prior work, flight distances were based on great circle route distance between the conference location and the home institution of paper authors.",
        "\n",
        "Previous estimates have not used flight connection data, which can help distinguish the effect of locating a conference near a well-connected international airport. Country-level geolocations are too coarse-grained, as many conference participants are from large countries like Canada, China, and the United States.\n",
        "\n",
        "The myclimate-based estimates include a multiplier to account for non-CO2 radiative forcing [[1](https://en.wikipedia.org/wiki/Environmental_effects_of_aviation#Climate_change), [2](https://www.myclimate.org/en/information/about-myclimate/downloads/flight-emission-calculator/)]. This is why per-participant emissions are larger in the studies that use this estimator. While it seems likely that CO2 emissions underestimate the climate impact of air travel, there is considerable uncertainty about the exact value of the multiplier (e.g., https://phys.org/news/2023-12-nasa-boeing-jet-contrails-science.html). For simplicity we examine only CO2 emissions, but we emphasize that the warming impacts of air travel are almost surely larger than the total emissions."
      ],
      "metadata": {
        "id": "n-dwibcgManP"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Code"
      ],
      "metadata": {
        "id": "kF_tET6nMnP1"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "mPVWwZW4oseV",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ef320b3c-5b6e-42d3-c795-1f85d7d6e033",
        "collapsed": true,
        "cellView": "form"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting haversine\n",
            "  Downloading haversine-2.8.0-py2.py3-none-any.whl (7.7 kB)\n",
            "Installing collected packages: haversine\n",
            "Successfully installed haversine-2.8.0\n",
            "Collecting fastparquet\n",
            "  Downloading fastparquet-2023.10.1-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.7 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.7/1.7 MB\u001b[0m \u001b[31m9.9 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: pandas>=1.5.0 in /usr/local/lib/python3.10/dist-packages (from fastparquet) (1.5.3)\n",
            "Requirement already satisfied: numpy>=1.20.3 in /usr/local/lib/python3.10/dist-packages (from fastparquet) (1.23.5)\n",
            "Collecting cramjam>=2.3 (from fastparquet)\n",
            "  Downloading cramjam-2.7.0-cp310-cp310-manylinux_2_17_x86_64.manylinux2014_x86_64.whl (1.6 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m1.6/1.6 MB\u001b[0m \u001b[31m18.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from fastparquet) (2023.6.0)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.10/dist-packages (from fastparquet) (23.2)\n",
            "Requirement already satisfied: python-dateutil>=2.8.1 in /usr/local/lib/python3.10/dist-packages (from pandas>=1.5.0->fastparquet) (2.8.2)\n",
            "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas>=1.5.0->fastparquet) (2023.3.post1)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.8.1->pandas>=1.5.0->fastparquet) (1.16.0)\n",
            "Installing collected packages: cramjam, fastparquet\n",
            "Successfully installed cramjam-2.7.0 fastparquet-2023.10.1\n"
          ]
        }
      ],
      "source": [
        "# @title imports\n",
        "\n",
        "from enum import Enum\n",
        "import matplotlib\n",
        "import matplotlib.pyplot as plt\n",
        "import networkx as nx\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "from scipy import spatial\n",
        "from scipy.spatial import cKDTree\n",
        "from google.colab import drive, files\n",
        "\n",
        "import logging\n",
        "logger = logging.getLogger('matplotlib.font_manager')\n",
        "logger.setLevel(logging.WARNING)\n",
        "\n",
        "!pip install haversine\n",
        "from haversine import haversine, Unit\n",
        "\n",
        "!pip install fastparquet  # for file io\n",
        "\n",
        "import os"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# @title constants\n",
        "HOME_DIR = 'drive/MyDrive/projects/acl_co2/'  # @param\n",
        "AIRPORT_FILENAME = 'airports.dat'  # @param\n",
        "ATTENDEE_FILENAME = 'acl_locations_with_lat_lon_fixed.csv'  # @param\n",
        "ROUTES_FILENAME = 'routes.dat'  # @param\n",
        "AIRPORT_DATA_PATH = os.path.join(HOME_DIR, AIRPORT_FILENAME)\n",
        "ATTENDEE_DATA_PATH = os.path.join(HOME_DIR, ATTENDEE_FILENAME)\n",
        "ROUTE_DATA_PATH = os.path.join(HOME_DIR, ROUTES_FILENAME)\n",
        "\n",
        "LOAD_SHORTEST_PATHS = True  # @param # For caching previous computations\n",
        "ASSUME_DIRECT = False  # @param        # Assume a direct flight exists between source and destination\n",
        "EARTH_RADIUS = 6371\n",
        "KG_CO2E_PER_PERSON = 4690  # @param # https://www.statista.com/statistics/268753/co2-emissions-per-capita-worldwide-since-1990/#:~:text=Global%20per%20capita%20carbon%20dioxide%20emissions%20averaged%204.69%20metric%20tons%20in%202021.\n",
        "KG_CO2E_PER_KWH = 0.436  # @param # https://ourworldindata.org/grapher/carbon-intensity-electricity?tab=chart\n",
        "A100_KW = 0.4  # @param\n",
        "MAX_GROUND_TRANSPORT_KM = 150  # @param\n",
        "NROWS = 100000 # debug\n",
        "\n",
        "# Two ways to compute CO2, based on myclimate and ICCT.\n",
        "Method = Enum('Method', ['ICCT', 'MYCLIMATE'])\n",
        "CO2_METHOD = Method.ICCT  # @param\n",
        "\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "plwhIhImNifR",
        "outputId": "c1d1a8c0-88e5-49b7-f68f-96bf73e2e103"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Loading data\n",
        "\n",
        "There are three data sources:\n",
        "\n",
        "- **Conference attendee locations**: City/state/nation data for each attendee.\n",
        "We wrote a script to geocode these locations, using the OpenStreetMaps provider in the python geocoder library (https://geocoder.readthedocs.io/providers/OpenStreetMap.html). That code is available on request.\n",
        "- **Airport locations**: Airport code and latitude/longitude, from https://github.com/jpatokal/openflights/tree/master/data.\n",
        "- **Flights**: All airport-to-airport flights, again from https://github.com/jpatokal/openflights/tree/master/data. Unfortunately this data hasn't been updated since 2014."
      ],
      "metadata": {
        "id": "AppDYliLyG81"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# @title data loading code\n",
        "def latlon_to_cartesian(lat, lon):\n",
        "  \"\"\"Converts latitude and longitude to cartesian coordinates for kdtree.\"\"\"\n",
        "  lat, lon = np.radians(lat), np.radians(lon)\n",
        "  x = np.cos(lat) * np.cos(lon)\n",
        "  y = np.cos(lat) * np.sin(lon)\n",
        "  z = np.sin(lat)\n",
        "  return x, y, z\n",
        "\n",
        "def read_attendees(attendee_path: str):\n",
        "  \"\"\"Read conference attendee file\"\"\"\n",
        "  names = ['count', 'city', 'state', 'country', 'lon', 'lat']\n",
        "\n",
        "  df_attendees = pd.read_csv(attendee_path, header=None, names=names, nrows=NROWS)\n",
        "\n",
        "  df_attendees['coordinates'] = df_attendees.apply(\n",
        "      lambda row: (float(row['lat']), float(row['lon'])), axis=1)\n",
        "\n",
        "  df_attendees['xyz'] = df_attendees.apply(\n",
        "    lambda row: latlon_to_cartesian(*row['coordinates']), axis=1)\n",
        "\n",
        "  total_attendees = df_attendees['count'].sum()\n",
        "  print(f\"Total attendees: {total_attendees}\")\n",
        "  print(f\"Total locations: {len(df_attendees)}\")\n",
        "\n",
        "  return df_attendees\n",
        "\n",
        "def load_all_data(airport_path: str = AIRPORT_DATA_PATH,\n",
        "                  routes_path: str = ROUTE_DATA_PATH,\n",
        "                  attendees_path: str = ATTENDEE_DATA_PATH):\n",
        "  df_airports = pd.read_csv(airport_path, header=None,\n",
        "        names=['name', 'city', 'country', 'code1', 'code2', 'lat', 'lon',\n",
        "              'code3', 'code4', 'code5', 'timezone', 'code7', 'code8'])\n",
        "\n",
        "  # clean up some tiny airports\n",
        "  df_airports = df_airports[df_airports.code1 != '\\\\N']\n",
        "  df_routes = pd.read_csv(routes_path, header=None, names=[\n",
        "      'airline', 'airline_id', 'source', 'source_id', 'destination',\n",
        "      'destination_id', 'codeshare', 'stops', 'equipment'])\n",
        "\n",
        "  # drop duplicate routes\n",
        "  df_routes = df_routes.drop_duplicates(subset=['source', 'destination']\n",
        "                                      ).reset_index(drop=True)\n",
        "  all_airports = sorted(\n",
        "        list(set(\n",
        "            df_routes['source'].unique()).union(df_routes['destination'].unique())))\n",
        "\n",
        "  df_airports = df_airports[\n",
        "      df_airports['code1'].isin(all_airports)].set_index('code1')\n",
        "  df_airports['coordinates'] = df_airports.apply(\n",
        "      lambda row: (row['lat'], row['lon']), axis=1)\n",
        "  # Cartesian coordinates for distance computations.\n",
        "  df_airports['xyz'] = df_airports.apply(\n",
        "      lambda row: latlon_to_cartesian(*row['coordinates']), axis=1)\n",
        "\n",
        "\n",
        "  df_attendees = read_attendees(attendees_path)\n",
        "  return df_airports, df_routes, df_attendees"
      ],
      "metadata": {
        "id": "Oq1WfRUjpbVq",
        "cellView": "form"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_airports, df_routes, df_attendees = load_all_data()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LdP7f-3nOCsX",
        "outputId": "34ca1852-f572-4603-96fe-266b072f8ebb"
      },
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Total attendees: 2588\n",
            "Total locations: 918\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Finding airports for each attendee location\n",
        "\n",
        "For each attendee, we find airports that they are likely to use. We do this in stages. If there are any airports within 75km, we include those. If not, we include all airports within 125km. If there are none, we include all airports within 250km. This covers >99% of attendees."
      ],
      "metadata": {
        "id": "9O7Xp04gyC1s"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# @title\n",
        "def add_attendee_airport_codes(df_attendees: pd.DataFrame,\n",
        "                               df_airports: pd.DataFrame) -> pd.DataFrame:\n",
        "  \"\"\"Uses KD tree to find closest airport code for each attendee.\"\"\"\n",
        "\n",
        "  # Build KDTree\n",
        "  kdtree = cKDTree(df_airports['xyz'].tolist())\n",
        "  def find_closest_airport_codes(\n",
        "      attendee: pd.Series,\n",
        "      radii: list[float] = [75/EARTH_RADIUS, 125/EARTH_RADIUS, 250/EARTH_RADIUS]):\n",
        "    \"\"\"Finds airports all within each radius of city.\"\"\"\n",
        "    for radius in radii:\n",
        "      locs = kdtree.query_ball_point(attendee['xyz'], r=radius)\n",
        "      candidates = df_airports.iloc[locs]\n",
        "      if len(candidates):\n",
        "        return candidates.index.tolist()\n",
        "    return None\n",
        "\n",
        "  return df_attendees.assign(airports=df_attendees.apply(\n",
        "        lambda row: find_closest_airport_codes(row), axis=1))\n",
        "\n",
        "df_attendees = add_attendee_airport_codes(df_attendees, df_airports)\n",
        "print(\"Random sample of cities and linked airports:\")\n",
        "print(df_attendees.query('count > 20').sample(5).reset_index(drop=True)[\n",
        "    ['city', 'state', 'country', 'airports']])"
      ],
      "metadata": {
        "id": "aRbkY5lkBus7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3dbbf4a0-3bba-4e51-b88b-363fc8948321"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Random sample of cities and linked airports:\n",
            "          city state        country                             airports\n",
            "0    vancouver    bc         canada  [BLI, YXX, YCD, YYJ, YVR, CXH, ESD]\n",
            "1  los_angeles    ca  united_states            [BUR, LAX, ONT, SNA, LGB]\n",
            "2    ann_arbor    mi  united_states                           [DTW, YQG]\n",
            "3      seattle    wa  united_states                           [BFI, SEA]\n",
            "4    sunnyvale    ca  united_states                      [SJC, OAK, SFO]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Shortest paths between all pairs of airports\n",
        "\n",
        "The main computational challenge is to find likely flights for each attendee. To do this, we build a graph of airports and then find the shortest paths between all pairs of nodes in the graph. Path length is defined first by the number of hops and then by the distance, so that a path with $n$ hops will be preferred over all paths with $n+1$ hops.\n",
        "\n",
        "This section also contains code for estimating CO2 emissions by flight distance, using two methods that reflect the higher emissions associated with takeoff: a piecewise linear function ([ICCT](https://theicct.org/sites/default/files/publications/ICCT_CO2-commercl-aviation-2018_20190918.pdf)) and a quadratic formula ([myclimate](https://www.myclimate.org/en/information/about-myclimate/downloads/flight-emission-calculator/))."
      ],
      "metadata": {
        "id": "8nnbkVPp0a-n"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# @title Shortest path and other code\n",
        "\n",
        "def co2_kg_by_distance_icct(distance: float) -> float:\n",
        "  \"\"\"Computes CO2 emissions by flight distance in KM based on the ICCT method\n",
        "  (https://theicct.org/sites/default/files/publications/ICCT_CO2-commercl-aviation-2018_20190918.pdf).\n",
        "  \"\"\"\n",
        "  co2_1 = min(distance, 500) * 0.16\n",
        "  co2_2 = max(0, min(distance-500, 500)) * 0.11\n",
        "  co2_3 = max(0, distance-1000) * 0.09\n",
        "  return co2_1 + co2_2 + co2_3\n",
        "\n",
        "def co2_kg_by_distance_myclimate(distance: float) -> float:\n",
        "  \"\"\"Computes co2 emissions by flight distance in KM based on the myclimate method\n",
        "  (https://www.myclimate.org/en/information/about-myclimate/downloads/flight-emission-calculator/).\n",
        "  \"\"\"\n",
        "  # x: Flight distance [km], defined as the sum of GCD (great circle distance)\n",
        "  #    and DC (distance correction for detours and holding patterns)\n",
        "  #    and inefficiencies in the air traffic control systems [km]\n",
        "  # S: Average number of seats (total across all cabin classes)\n",
        "  # PLF: Passenger load factor\n",
        "  # CF: Cargo factor\n",
        "  # CW: Cabin class weighting factor\n",
        "  # EF: CO₂ emission factor for jet fuel combustion (kerosene)\n",
        "  # M: Multiplier accounting for potential non-CO₂ effects\n",
        "  # P: CO₂e emission factor for pre-production jet fuel, kerosene\n",
        "  # AF: Aircraft factor\n",
        "  # A: Airport infrastructure emissions\n",
        "\n",
        "  aircrafts = ['Standard short-haul flight', 'Standard long-haul flight']\n",
        "  DC = 95\n",
        "\n",
        "  x = distance+ DC\n",
        "\n",
        "  S = [157.86, 302.58]\n",
        "  PLF = [.796, 0.82]\n",
        "  CF = 2*[0.26]\n",
        "  CW = 2*[1]\n",
        "  EF = 2*[3.16]\n",
        "  P = 2*[0.538]\n",
        "  M = 2*[3]\n",
        "  AF = 2*[0.00034]\n",
        "  A = 2*[11.68]\n",
        "  a = [0.000007, 0.00029]\n",
        "  b = [2.775, 3.475]\n",
        "  c = [1260.608, 3259.691]\n",
        "\n",
        "  MAX_SHORT = 1500\n",
        "  MIN_LONG = 2500\n",
        "  def co2(x, S, PLF, CF, CW, EF, P, M, AF, A, a, b, c):\n",
        "    return (a * x**2+b*x+c)/(S*PLF) *(1-CF) * CW*(EF*M+P) + AF*x+A\n",
        "\n",
        "  if distance <= MAX_SHORT:\n",
        "    index = 0\n",
        "    return co2(x, S[index], PLF[index], CF[index], CW[index],\n",
        "               EF[index], P[index], M[index], AF[index], A[index],\n",
        "               a[index], b[index], c[index])\n",
        "  elif distance >= MIN_LONG:\n",
        "    index = 1\n",
        "    return co2(x, S[index], PLF[index], CF[index], CW[index],\n",
        "               EF[index], P[index], M[index], AF[index], A[index],\n",
        "               a[index], b[index], c[index])\n",
        "  else:\n",
        "    index = 0\n",
        "    co2_1 = co2(x, S[index], PLF[index], CF[index], CW[index],\n",
        "               EF[index], P[index], M[index], AF[index], A[index],\n",
        "               a[index], b[index], c[index])\n",
        "\n",
        "    index = 1\n",
        "    co2_2 = co2(x, S[index], PLF[index], CF[index], CW[index],\n",
        "               EF[index], P[index], M[index], AF[index], A[index],\n",
        "               a[index], b[index], c[index])\n",
        "\n",
        "    coefficient = (distance-MAX_SHORT)/(MIN_LONG-MAX_SHORT)\n",
        "\n",
        "  return coefficient*co2_1 + (1-coefficient)*co2_2\n",
        "\n",
        "\n",
        "def build_shortest_paths(df_airports: pd.DataFrame,\n",
        "                         df_routes: pd.DataFrame) -> pd.DataFrame:\n",
        "  \"\"\"Build a dataframe of paths and lengths between pairs of airports.\"\"\"\n",
        "  G = nx.Graph()\n",
        "\n",
        "  for code, airport in df_airports.iterrows():\n",
        "      G.add_node(code, pos=airport['coordinates'])\n",
        "\n",
        "  # Calculate the Haversine distances between airports and add weighted edges to\n",
        "  # the graph\n",
        "  for _, route in df_routes.iterrows():\n",
        "    source = route['source']\n",
        "    destination = route['destination']\n",
        "    if source in df_airports.index and destination in df_airports.index:\n",
        "      coordinates_source = df_airports.loc[source]['coordinates']\n",
        "      coordinates_destination = df_airports.loc[destination]['coordinates']\n",
        "      distance = haversine(coordinates_source, coordinates_destination,\n",
        "                          unit=Unit.KILOMETERS)\n",
        "      G.add_edge(source, destination, weight=distance,\n",
        "                 weight_with_hops=(1 + distance/1e5),\n",
        "                 kg_co2e_icct=co2_kg_by_distance_icct(distance),\n",
        "                 kg_co2e_myclimate=co2_kg_by_distance_myclimate(distance))\n",
        "\n",
        "  # Use NetworkX's all_pairs_shortest_path to find the shortest paths between\n",
        "  # all pairs of airports.\n",
        "  # weight_with_hops will choose the shortest combined distance of the\n",
        "  # fewest-hop routes\n",
        "  shortest_paths = dict(nx.all_pairs_dijkstra_path(G, weight='weight_with_hops'))\n",
        "  shortest_path_lengths = dict(nx.all_pairs_dijkstra_path_length(G))\n",
        "  shortest_path_co2_icct = dict(nx.all_pairs_dijkstra_path_length(G, weight='kg_co2e_icct'))\n",
        "  shortest_path_co2_myclimate = dict(nx.all_pairs_dijkstra_path_length(G, weight='kg_co2e_myclimate'))\n",
        "\n",
        "  records = []\n",
        "  for source in shortest_paths.keys():\n",
        "    for destination in shortest_paths[source].keys():\n",
        "      records.append({\"source\": source, \"destination\": destination,\n",
        "                      \"path\": shortest_paths[source][destination],\n",
        "                      \"distance\": shortest_path_lengths[source][destination],\n",
        "                      \"kg_co2e_icct\": shortest_path_co2_icct[source][destination],\n",
        "                      \"kg_co2e_myclimate\": shortest_path_co2_myclimate[source][destination]})\n",
        "  return pd.DataFrame(records)\n",
        "\n",
        "def find_best_paths_per_attendee(\n",
        "    df_attendees: pd.DataFrame,\n",
        "    paths_to_target: pd.DataFrame,\n",
        "    max_ground_travel: float = MAX_GROUND_TRANSPORT_KM) -> pd.DataFrame:\n",
        "  \"\"\"Finds best paths to the target destination per attendee.\"\"\"\n",
        "  df_per_attendee_airport = pd.merge(\n",
        "      df_attendees.explode('airports').reset_index(),\n",
        "      paths_to_target,\n",
        "      left_on='airports', right_on='source')\n",
        "  best_rows = df_per_attendee_airport.groupby('index')['distance'].idxmin()\n",
        "  df_best_paths = df_per_attendee_airport.loc[best_rows]\n",
        "\n",
        "  target = paths_to_target.destination.unique()[0]\n",
        "  ground_distances = df_best_paths['coordinates'].apply(\n",
        "          lambda x: haversine(x, df_airports.loc[target]['coordinates'],\n",
        "                              unit=Unit.KILOMETERS))\n",
        "  df_best_paths['ground_distance'] = ground_distances\n",
        "\n",
        "  ground_travel = df_best_paths['ground_distance'] < max_ground_travel\n",
        "  df_best_paths.loc[ground_travel, 'num_hops'] = 0\n",
        "  df_best_paths.loc[ground_travel, 'kg_co2e_icct'] = 0\n",
        "  df_best_paths.loc[ground_travel, 'kg_co2e_myclimate'] = 0\n",
        "  df_best_paths.loc[ground_travel, 'distance'] = 0\n",
        "  df_best_paths['path'] = df_best_paths.apply(\n",
        "      lambda row: [target] if row['ground_distance'] < max_ground_travel \\\n",
        "      else row['path'], axis=1)\n",
        "  return df_best_paths\n",
        "\n",
        "\n",
        "def find_best_direct_path_per_attendee(\n",
        "    df_attendees: pd.DataFrame,\n",
        "    target: str,\n",
        "    target_coordinates,\n",
        "    max_ground_travel: float = MAX_GROUND_TRANSPORT_KM) -> pd.DataFrame:\n",
        "  \"\"\"\n",
        "  An ablation study:\n",
        "  Compute direct distance to target destination per attendee,\n",
        "  instead of using flight information.\n",
        "  Use ground distance as a proxy (not taking airport locations into account)\n",
        "  \"\"\"\n",
        "  ground_distances = df_attendees['coordinates'].apply(\n",
        "          lambda row: haversine(row, target_coordinates, unit=Unit.KILOMETERS))\n",
        "\n",
        "  df_best_paths = pd.DataFrame()\n",
        "  df_best_paths['distance'] = ground_distances\n",
        "  df_best_paths['kg_co2e_icct'] = ground_distances.apply(\n",
        "      lambda x: co2_kg_by_distance_icct(x))\n",
        "  df_best_paths['kg_co2e_myclimate'] = ground_distances.apply(\n",
        "      lambda x: co2_kg_by_distance_myclimate(x))\n",
        "\n",
        "  # For distance shorter than max_ground_travel, assume attendees didn't fly,\n",
        "  # and assuming 0 kg_co2e.\n",
        "  ground_travel = df_best_paths['distance'] < max_ground_travel\n",
        "  df_best_paths.loc[ground_travel, 'num_hops'] = 0\n",
        "  df_best_paths.loc[ground_travel, 'kg_co2e_icct'] = 0\n",
        "  df_best_paths.loc[ground_travel, 'kg_co2e_myclimate'] = 0\n",
        "  df_best_paths.loc[ground_travel, 'distance'] = 0\n",
        "\n",
        "  air_travel = df_best_paths['distance'] >= max_ground_travel\n",
        "  df_best_paths.loc[air_travel, 'num_hops'] = 1\n",
        "\n",
        "  df_best_paths['path'] = df_attendees.apply(\n",
        "      lambda row: [str(row['city'])+','+str(row['state'])+','+str(row['country']), target], axis=1)\n",
        "\n",
        "  df_best_paths['count'] = df_attendees.apply(\n",
        "      lambda row: row['count'], axis=1)\n",
        "\n",
        "  df_best_paths['path'] = df_best_paths.apply(\n",
        "      lambda row: target if row['distance'] < max_ground_travel \\\n",
        "      else row['path'], axis=1)\n",
        "  return df_best_paths\n",
        "\n",
        "\n",
        "def analyze_best_paths(df_best_paths: pd.DataFrame, df_attendees: pd.DataFrame,\n",
        "                       co2_method=CO2_METHOD):\n",
        "  \"\"\"Some helper code to print out details of the best paths.\"\"\"\n",
        "  pct_accounted_for = df_best_paths['count'].sum() / df_attendees['count'].sum()\n",
        "  counts = df_best_paths['count'].to_numpy()\n",
        "  distances = df_best_paths['distance'].to_numpy()\n",
        "  co2_method_str = \"icct\" if co2_method==Method.ICCT else \"myclimate\"\n",
        "  kg_co2e = df_best_paths[f'kg_co2e_{co2_method_str}'].to_numpy()\n",
        "\n",
        "  # A factor 2 is added to both distance and co2e to account for roundtrips\n",
        "  total_distance = 2 * counts @ distances\n",
        "  kg_co2e = 2 * kg_co2e @ counts\n",
        "\n",
        "  people_years = kg_co2e / KG_CO2E_PER_PERSON\n",
        "  a100_co2_per_hour = KG_CO2E_PER_KWH * A100_KW\n",
        "  a100_hours = kg_co2e / a100_co2_per_hour\n",
        "\n",
        "  direct_flights = np.array(df_best_paths['num_hops'] <= 1)\n",
        "  pct_direct_flights = (counts @ direct_flights) / counts.sum()\n",
        "\n",
        "  ground_travel = np.array(df_best_paths['num_hops'] == 0)\n",
        "  pct_ground_travel = (counts @ ground_travel) / counts.sum()\n",
        "\n",
        "  return {\"kilometers flight\": total_distance,\n",
        "          \"mt co2e\": kg_co2e / 1000,\n",
        "          \"a100 hours\": a100_hours,\n",
        "          \"a100 years\": a100_hours / (365 * 24),\n",
        "          \"person_years_co2e\": people_years,\n",
        "          \"pct direct flights\": pct_direct_flights,\n",
        "          \"pct ground travel\": pct_ground_travel,\n",
        "          \"pct accounted for:\": pct_accounted_for}\n",
        "\n",
        "def print_path_stats(paths, df_attendees, co2_method: str = 'icct'):\n",
        "  for key, val in analyze_best_paths(paths, df_attendees,\n",
        "                                     co2_method=co2_method).items():\n",
        "    if val > 10000:\n",
        "      val_str = f\"{val:.2e}\"\n",
        "    elif val > 1:\n",
        "      val_str = f\"{val:.1f}\"\n",
        "    else:\n",
        "      val_str = f\"{val:.3f}\"\n",
        "    print(f\"{key}: {val_str}\")"
      ],
      "metadata": {
        "id": "64txKHxG14Ll",
        "cellView": "form"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "LOAD_SHORTEST_PATHS = True"
      ],
      "metadata": {
        "id": "RlnpaILp8d6-"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Load or compute best paths\n",
        "shortest_path_path = os.path.join(HOME_DIR, \"shortest_paths.parquet\")\n",
        "if LOAD_SHORTEST_PATHS:\n",
        "  df_paths = pd.read_parquet(shortest_path_path,\n",
        "                             engine=\"fastparquet\")\n",
        "else:\n",
        "  df_paths = build_shortest_paths(df_airports, df_routes)  # this takes ~20 minutes\n",
        "  df_paths['num_hops'] = df_paths['path'].apply(len) - 1\n",
        "  df_paths.to_parquet(shortest_path_path, engine=\"fastparquet\")"
      ],
      "metadata": {
        "id": "6d38f_rQ6Xfh"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Emissions estimate\n",
        "\n",
        "Finally we can find the shortest path to the conference venue for each attendee. We make the conservative assumption that everyone located within 150km of Toronto emitted zero carbon while traveling to the conference."
      ],
      "metadata": {
        "id": "nKWl6dzh9-aM"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Print best paths\n",
        "best_paths = find_best_paths_per_attendee(\n",
        "    df_attendees,\n",
        "    df_paths[df_paths['destination']=='YYZ'],\n",
        ")\n",
        "\n",
        "print_path_stats(best_paths, df_attendees, co2_method=Method.ICCT)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "soe4XE_wA4M_",
        "outputId": "514657a0-100b-484f-b4ad-b757bdd021ef"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "kilometers flight: 2.36e+07\n",
            "mt co2e: 2402.3\n",
            "a100 hours: 1.38e+07\n",
            "a100 years: 1572.5\n",
            "person_years_co2e: 512.2\n",
            "pct direct flights: 0.587\n",
            "pct ground travel: 0.089\n",
            "pct accounted for:: 1.000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# myclimate estimate\n",
        "print_path_stats(best_paths, df_attendees, co2_method=Method.MYCLIMATE)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mFGeYR57GEa9",
        "outputId": "49c1aa91-bd2a-4c95-ca93-8c2994b99018"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "kilometers flight: 2.36e+07\n",
            "mt co2e: 4662.6\n",
            "a100 hours: 2.67e+07\n",
            "a100 years: 3051.9\n",
            "person_years_co2e: 994.2\n",
            "pct direct flights: 0.587\n",
            "pct ground travel: 0.089\n",
            "pct accounted for:: 1.000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The myclimate estimate is nearly two times higher. As noted above, the main reason for this difference is that myclimate incorporates a large multiplier for non-co2 warming effects."
      ],
      "metadata": {
        "id": "VqoHZvWVGJ-T"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Ablation\n",
        "\n",
        "One distinct aspect of our methodology is the use of flight routes rather than direct city-to-city distances. We test the effect of this on our overall emissions estimate."
      ],
      "metadata": {
        "id": "NUGBxGPFF2Z4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# @title\n",
        "def flight_path_ablation(target: str):\n",
        "  \"\"\"Ablate the effect of modeling flight paths.\"\"\"\n",
        "  best_paths = find_best_paths_per_attendee(\n",
        "      df_attendees,\n",
        "      df_paths[df_paths['destination']==target]\n",
        "  )\n",
        "\n",
        "  best_path_stats = analyze_best_paths(best_paths, df_attendees)\n",
        "  direct_paths = find_best_direct_path_per_attendee(\n",
        "      df_attendees, target, df_airports.loc[target].coordinates)\n",
        "  direct_path_stats = analyze_best_paths(direct_paths, df_attendees)\n",
        "\n",
        "  best_path_co2 = best_path_stats['mt co2e']\n",
        "  direct_path_co2 = direct_path_stats['mt co2e']\n",
        "\n",
        "  pct_connecting = (1 - best_path_stats['pct direct flights'])*100\n",
        "  co2_increase = (1 - direct_path_co2/best_path_co2)*100\n",
        "  print(f\"{co2_increase:.1f}% increase in emissions from {pct_connecting:.1f}%\"\\\n",
        "        f\" of attendees taking connecting flights to {target} \"\\\n",
        "        f\"({direct_path_co2:.0f} to {best_path_co2:.0f} mt co2e).\")"
      ],
      "metadata": {
        "id": "JoZsWc8hEfMG"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "flight_path_ablation('YYZ')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4Bc2gGz9YqqD",
        "outputId": "9f0a1b7e-302f-483f-bf93-57893ae1c259"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "4.1% increase in emissions from 41.3% of attendees taking connecting flights to YYZ (2305 to 2402 mt co2e).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "The difference is relatively small, but this is in part because Toronto is a well-connected airport with a relatively high proportion of direct flights (59\\%). The impact of modeling flight routes is greater if we pick a less well-connected city such as Montreal (37.5\\% direct flights):"
      ],
      "metadata": {
        "id": "Z97gWw3bXGXR"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "flight_path_ablation('YUL')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VGO20zYgYtLn",
        "outputId": "f1dd9c9c-015f-4904-c906-b448627c392e"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "7.8% increase in emissions from 62.5% of attendees taking connecting flights to YUL (2326 to 2523 mt co2e).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "For Montreal, modeling flight connections leads to a 7.8\\% increase in the CO2 estimate. In reality, the increase is likely to be higher, because we have assumed that each attendee chose the most carbon-efficient flight path.\n",
        "\n",
        "Note that these estimates assume the same geographic distribution of participants regardless of where the conference was held, so the absolute emissions numbers should not be considered to be accurate."
      ],
      "metadata": {
        "id": "_lnGWDFeZ7Ra"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Limitations\n",
        "\n",
        "- We don't actually know how people got to ACL, we can only make an educated guess based on their addresses and commercial flight routes.\n",
        "- The flight routes are based on data from 2014. If there are more direct flights now, the total emissions would be somewhat lower. On the other hand, we assume maximally carbon-efficient flight connections, which underestimates emissions.\n",
        "- We don't incorporate differences between airplanes. For example, flights from remote airports might be on smaller planes with higher per-passenger emissions.\n",
        "- The A100 estimate (6.9 million hours) is based on the global average greenhouse gas intensity of electricity production. Many data centers are powered by cleaner sources of energy. ACL 2023's A100-equivalent would be four times greater if we used the much lower greenhouse gas intensity of electricity production in Jacob's home state of Washington (https://www.washingtonpost.com/climate-environment/interactive/2023/clean-energy-electricity-sources/).\n",
        "- If people didn't go to ACL, maybe they would have flown somewhere else. For example, maybe they would have visited Toronto for vacation (it was very fun!), or maybe they would have gone to a different conference.\n",
        "- Canceling ACL would not have eliminated all of these emissions because most of the flights would have happened even if there were a few more empty seats.\n",
        "\n",
        "The last two points are counterfactuals, and whether they are a helpful way to think about emissions reductions is a philosophical question that is beyond the scope of this colab. It seems uncontroversial that a broad-based reduction in conference travel (e.g., across the computer science research community) could have a significant impact.\n",
        "\n",
        "### Comparison with LLMs\n",
        "\n",
        "Time to address the elephant (llama?) in the room. The ACL 2023 travel emissions are on a similar scale to the numbers reported in papers that quantify the emissions associated with training some of the largest language models. However, LLM training emissions are increasingly mitigated by the transition to low-carbon sources of energy. For commercial air travel, the situation is different: we may be decades away from alternatives to fossil fuels. Until then, flying less is the only way to reduce travel-related emissions.\n"
      ],
      "metadata": {
        "id": "vm7QNsTJ9ij9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Bonus visualization"
      ],
      "metadata": {
        "id": "VRAbSM-OZO14"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install basemap"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 516
        },
        "id": "AeNFu45yZR-n",
        "outputId": "1419c324-0ed2-467c-cb59-7d1b3b59dbbd"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting basemap\n",
            "  Downloading basemap-1.3.8-cp310-cp310-manylinux1_x86_64.whl (860 kB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m860.7/860.7 kB\u001b[0m \u001b[31m4.4 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hCollecting basemap-data<1.4,>=1.3.2 (from basemap)\n",
            "  Downloading basemap_data-1.3.2-py2.py3-none-any.whl (30.5 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m30.5/30.5 MB\u001b[0m \u001b[31m19.6 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: pyshp<2.4,>=1.2 in /usr/local/lib/python3.10/dist-packages (from basemap) (2.3.1)\n",
            "Requirement already satisfied: matplotlib<3.8,>=1.5 in /usr/local/lib/python3.10/dist-packages (from basemap) (3.7.1)\n",
            "Requirement already satisfied: pyproj<3.7.0,>=1.9.3 in /usr/local/lib/python3.10/dist-packages (from basemap) (3.6.1)\n",
            "Requirement already satisfied: numpy<1.26,>=1.21 in /usr/local/lib/python3.10/dist-packages (from basemap) (1.23.5)\n",
            "Requirement already satisfied: contourpy>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib<3.8,>=1.5->basemap) (1.2.0)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.10/dist-packages (from matplotlib<3.8,>=1.5->basemap) (0.12.1)\n",
            "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib<3.8,>=1.5->basemap) (4.45.1)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib<3.8,>=1.5->basemap) (1.4.5)\n",
            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib<3.8,>=1.5->basemap) (23.2)\n",
            "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.10/dist-packages (from matplotlib<3.8,>=1.5->basemap) (9.4.0)\n",
            "Requirement already satisfied: pyparsing>=2.3.1 in /usr/local/lib/python3.10/dist-packages (from matplotlib<3.8,>=1.5->basemap) (3.1.1)\n",
            "Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.10/dist-packages (from matplotlib<3.8,>=1.5->basemap) (2.8.2)\n",
            "Requirement already satisfied: certifi in /usr/local/lib/python3.10/dist-packages (from pyproj<3.7.0,>=1.9.3->basemap) (2023.11.17)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.7->matplotlib<3.8,>=1.5->basemap) (1.16.0)\n",
            "Installing collected packages: basemap-data, basemap\n",
            "Successfully installed basemap-1.3.8 basemap-data-1.3.2\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.colab-display-data+json": {
              "pip_warning": {
                "packages": [
                  "mpl_toolkits"
                ]
              }
            }
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "from mpl_toolkits.basemap import Basemap"
      ],
      "metadata": {
        "id": "yMQv9n4R3iNO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "top_sources = df_attendees.sort_values('count', ascending=False).loc[1:7]\n",
        "city_list = top_sources['city'].to_list()\n",
        "coordinates_list = top_sources['coordinates'].to_list()\n",
        "count_list = top_sources['count'].to_list()\n",
        "city_counts = df_attendees.set_index('city')['count'].to_dict()"
      ],
      "metadata": {
        "id": "_6mVNFIm41qv"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# @title\n",
        "# Convert Cartesian coordinates back to geographical coordinates\n",
        "def convert_to_geographical(x, y, z):\n",
        "    lat_rad = np.arcsin(z)\n",
        "    lon_rad = np.arctan2(y, x)\n",
        "    lat = np.degrees(lat_rad)\n",
        "    lon = np.degrees(lon_rad)\n",
        "    return lat, lon\n",
        "\n",
        "toronto_lat, toronto_lon = 43.70, -79.42\n",
        "destinations = {city_i: coordinates_i for city_i, coordinates_i in\n",
        "                zip(city_list, coordinates_list)}\n",
        "# Find center of bounding box containing all destinations\n",
        "locs_xyz = [latlon_to_cartesian(*coords) for coords in\n",
        "            list(destinations.values()) + [(toronto_lat, toronto_lon)]]\n",
        "bb_center = .5 * (np.array(locs_xyz).max(0) + np.array(locs_xyz).min(0))\n",
        "bb_center_latlon = convert_to_geographical(*np.array(locs_xyz).mean(0))\n",
        "# bb_center_latlon = convert_to_geographical(*bb_center)\n",
        "\n",
        "# Create a Basemap with an Orthographic Projection centered on Toronto\n",
        "m = Basemap(\n",
        "    projection='ortho',\n",
        "    lat_0=toronto_lat + 20,  # bb_center_latlon[0], # + 25,\n",
        "    lon_0=toronto_lon + 5,  # bb_center_latlon[1], # + 10,  # avoid this hack\n",
        "    resolution='c',\n",
        ")\n",
        "\n",
        "# Draw coastlines and countries\n",
        "coastlines = m.drawcoastlines()\n",
        "coastlines.set_alpha(0.5)\n",
        "countries = m.drawcountries()\n",
        "countries.set_alpha(0.4)\n",
        "\n",
        "# Plot Toronto\n",
        "x_toronto, y_toronto = m(toronto_lon, toronto_lat)\n",
        "# m.plot(x_toronto, y_toronto, 'bo', markersize=8, label='Toronto')\n",
        "\n",
        "# Plot destinations and draw great circle routes\n",
        "for city, (lat, lon) in destinations.items():\n",
        "    x_city, y_city = m(lon, lat)\n",
        "    m.drawgreatcircle(toronto_lon, toronto_lat, lon, lat,\n",
        "                      linewidth=1 + 2 * city_counts[city] / 100,\n",
        "                      color='r', linestyle='-', alpha=0.8)\n",
        "    m.plot(x_city, y_city, 'g*', markersize=5, label=city)\n",
        "    if city == \"new_york\":\n",
        "      ha='left'\n",
        "    else:\n",
        "      ha='right'\n",
        "    plt.text(x_city, y_city, city.title().replace('_', ' '), fontsize=9,\n",
        "             ha=ha, va='bottom', color='black', weight='bold')\n",
        "\n",
        "# Add a title\n",
        "plt.title('Great Circle Routes from Toronto to Top Attendee Addresses')\n",
        "\n",
        "# Show the map\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "FBKFWEKU3bKr",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 428
        },
        "cellView": "form",
        "outputId": "f5a213bb-84ba-4ff8-8fad-a293b399d952"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Acknowledgements\n",
        "\n",
        "We are grateful to the ACL for providing the aggregated conference participation data, and to Sasha Luccioni for her feedback."
      ],
      "metadata": {
        "id": "IlFxwbJF7S8B"
      }
    }
  ]
}