](https://colab.research.google.com/drive/1PAOktBqh_3QzgAKi53F4JbQxoOuBsUBY?usp=sharing) | How to load and explore RelBench data
-| Training models | [](https://colab.research.google.com/drive/1_z0aKcs5XndEacX1eob6csDuR4DYhGQU?usp=sharing)| Train your first GNN-based model on RelBench. |
+| Name | Notebook | Try on Colab | Description |
+|-------|----------|--------------|---------------------------------------------------------|
+| Loading Data | [load_data.ipynb](tutorials/load_data.ipynb) | [](https://colab.research.google.com/github/snap-stanford/relbench/blob/main/tutorials/load_data.ipynb) | How to load and explore RelBench data
+| Training models | [train_model.ipynb](tutorials/train_model.ipynb) | [](https://colab.research.google.com/github/snap-stanford/relbench/blob/main/tutorials/train_model.ipynb)| Train your first GNN-based model on RelBench. |
diff --git a/tutorials/load_data.ipynb b/tutorials/load_data.ipynb
new file mode 100644
index 00000000..23067119
--- /dev/null
+++ b/tutorials/load_data.ipynb
@@ -0,0 +1,1259 @@
+{
+ "nbformat": 4,
+ "nbformat_minor": 0,
+ "metadata": {
+ "colab": {
+ "provenance": []
+ },
+ "kernelspec": {
+ "name": "python3",
+ "display_name": "Python 3"
+ },
+ "language_info": {
+ "name": "python"
+ }
+ },
+ "cells": [
+ {
+ "cell_type": "markdown",
+ "source": [
+ "First, `pip install`"
+ ],
+ "metadata": {
+ "id": "9hVIN7QW2etU"
+ }
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "QkzAsRon2ZIw",
+ "outputId": "4dca4023-9f6b-4124-a629-3b9257844ba8"
+ },
+ "outputs": [
+ {
+ "output_type": "stream",
+ "name": "stdout",
+ "text": [
+ "Collecting relbench==1.0.0rc1\n",
+ " Downloading relbench-1.0.0rc1-py3-none-any.whl (57 kB)\n",
+ "\u001b[?25l \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m0.0/57.4 kB\u001b[0m \u001b[31m?\u001b[0m eta \u001b[36m-:--:--\u001b[0m\r\u001b[2K \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m57.4/57.4 kB\u001b[0m \u001b[31m1.5 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
+ "\u001b[?25hRequirement already satisfied: pandas in /usr/local/lib/python3.10/dist-packages (from relbench==1.0.0rc1) (2.0.3)\n",
+ "Requirement already satisfied: pooch in /usr/local/lib/python3.10/dist-packages (from relbench==1.0.0rc1) (1.8.2)\n",
+ "Requirement already satisfied: pyarrow in /usr/local/lib/python3.10/dist-packages (from relbench==1.0.0rc1) (14.0.2)\n",
+ "Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from relbench==1.0.0rc1) (1.25.2)\n",
+ "Requirement already satisfied: duckdb in /usr/local/lib/python3.10/dist-packages (from relbench==1.0.0rc1) (0.10.3)\n",
+ "Requirement already satisfied: scikit-learn in /usr/local/lib/python3.10/dist-packages (from relbench==1.0.0rc1) (1.2.2)\n",
+ "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.10/dist-packages (from relbench==1.0.0rc1) (4.12.2)\n",
+ "Requirement already satisfied: tqdm in /usr/local/lib/python3.10/dist-packages (from relbench==1.0.0rc1) (4.66.4)\n",
+ "Requirement already satisfied: python-dateutil>=2.8.2 in /usr/local/lib/python3.10/dist-packages (from pandas->relbench==1.0.0rc1) (2.8.2)\n",
+ "Requirement already satisfied: pytz>=2020.1 in /usr/local/lib/python3.10/dist-packages (from pandas->relbench==1.0.0rc1) (2023.4)\n",
+ "Requirement already satisfied: tzdata>=2022.1 in /usr/local/lib/python3.10/dist-packages (from pandas->relbench==1.0.0rc1) (2024.1)\n",
+ "Requirement already satisfied: platformdirs>=2.5.0 in /usr/local/lib/python3.10/dist-packages (from pooch->relbench==1.0.0rc1) (4.2.2)\n",
+ "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from pooch->relbench==1.0.0rc1) (24.1)\n",
+ "Requirement already satisfied: requests>=2.19.0 in /usr/local/lib/python3.10/dist-packages (from pooch->relbench==1.0.0rc1) (2.31.0)\n",
+ "Requirement already satisfied: scipy>=1.3.2 in /usr/local/lib/python3.10/dist-packages (from scikit-learn->relbench==1.0.0rc1) (1.11.4)\n",
+ "Requirement already satisfied: joblib>=1.1.1 in /usr/local/lib/python3.10/dist-packages (from scikit-learn->relbench==1.0.0rc1) (1.4.2)\n",
+ "Requirement already satisfied: threadpoolctl>=2.0.0 in /usr/local/lib/python3.10/dist-packages (from scikit-learn->relbench==1.0.0rc1) (3.5.0)\n",
+ "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.10/dist-packages (from python-dateutil>=2.8.2->pandas->relbench==1.0.0rc1) (1.16.0)\n",
+ "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests>=2.19.0->pooch->relbench==1.0.0rc1) (3.3.2)\n",
+ "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests>=2.19.0->pooch->relbench==1.0.0rc1) (3.7)\n",
+ "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests>=2.19.0->pooch->relbench==1.0.0rc1) (2.0.7)\n",
+ "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests>=2.19.0->pooch->relbench==1.0.0rc1) (2024.6.2)\n",
+ "Installing collected packages: relbench\n",
+ "Successfully installed relbench-1.0.0rc1\n"
+ ]
+ }
+ ],
+ "source": [
+ "!pip install relbench==1.0.0rc1"
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "To start we can check all of the databases currently available in RelBench by printing:"
+ ],
+ "metadata": {
+ "id": "yLfAfxNO-LBV"
+ }
+ },
+ {
+ "cell_type": "code",
+ "source": [
+ "from relbench.datasets import get_dataset_names\n",
+ "\n",
+ "get_dataset_names()"
+ ],
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "ApqEZu5--HJy",
+ "outputId": "f45a7623-ce90-4dd9-c088-004290d6519a"
+ },
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "['rel-amazon',\n",
+ " 'rel-avito',\n",
+ " 'rel-event',\n",
+ " 'rel-f1',\n",
+ " 'rel-hm',\n",
+ " 'rel-stack',\n",
+ " 'rel-trial']"
+ ]
+ },
+ "metadata": {},
+ "execution_count": 2
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "# Get dataset\n",
+ "\n",
+ "Let's start with the F1 dataset since it's the smallest and is easy to work with. All it takes is one line!\n"
+ ],
+ "metadata": {
+ "id": "foOR0gDJ2j1s"
+ }
+ },
+ {
+ "cell_type": "code",
+ "source": [
+ "from relbench.datasets import get_dataset\n",
+ "\n",
+ "dataset = get_dataset(name=\"rel-f1\", download=True)"
+ ],
+ "metadata": {
+ "id": "RsAPkjOk2hDn",
+ "outputId": "837cbd40-167b-4bb1-968e-7120de1fa028",
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ }
+ },
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "stream",
+ "name": "stderr",
+ "text": [
+ "Downloading file 'rel-f1/db.zip' from 'https://relbench.stanford.edu/download/rel-f1/db.zip' to '/root/.cache/relbench'.\n",
+ "100%|████████████████████████████████████████| 704k/704k [00:00<00:00, 174MB/s]\n",
+ "Unzipping contents of '/root/.cache/relbench/rel-f1/db.zip' to '/root/.cache/relbench/rel-f1/.'\n"
+ ]
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "Use `download=True` the first time you load a patricular dataset to automatically download the data from the RelBench server onto your machine.\n",
+ "\n",
+ "Now we have loaded the database, let's start poking around to see what's inside."
+ ],
+ "metadata": {
+ "id": "M3dJVd8A2sOt"
+ }
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "# Val / Test cutoffs\n",
+ "\n",
+ "We can check the val/test time cutoffs as follows:"
+ ],
+ "metadata": {
+ "id": "lgeqc7x82596"
+ }
+ },
+ {
+ "cell_type": "code",
+ "source": [
+ "dataset.val_timestamp, dataset.test_timestamp"
+ ],
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "239e0MQB2yMO",
+ "outputId": "07b2cf05-a8d4-4c6c-b5cd-a755617a5baa"
+ },
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "(Timestamp('2005-01-01 00:00:00'), Timestamp('2010-01-01 00:00:00'))"
+ ]
+ },
+ "metadata": {},
+ "execution_count": 4
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "This means that information upto 2005 can be used for training, and upto 2010 can be used for validation.\n",
+ "\n",
+ "Note that it is a RelBench design choice to make the validation and test cutoffs a dataset property, _not_ a task-specific property. In other words, all tasks for a given database use the same time splits.\n",
+ "\n",
+ "\n",
+ "# Load database\n",
+ "\n",
+ "Next we check out the database itself..."
+ ],
+ "metadata": {
+ "id": "-75pIuTV3Ae6"
+ }
+ },
+ {
+ "cell_type": "code",
+ "source": [
+ "db = dataset.get_db()"
+ ],
+ "metadata": {
+ "id": "dm1f8YDF273p"
+ },
+ "execution_count": null,
+ "outputs": []
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "*This* returns a RelBench `Database` object. By default, the rows with timestamp > test_timestamp are excluded to prevent accidental test set leakage.\n",
+ "The complete database can be loaded with `database.get_db(upto_test_timestamp=False)`."
+ ],
+ "metadata": {
+ "id": "kpBRvvqb3FXQ"
+ }
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "With this we can double check the full timespan of the database:"
+ ],
+ "metadata": {
+ "id": "H9mRPjk33JfQ"
+ }
+ },
+ {
+ "cell_type": "code",
+ "source": [
+ "db.min_timestamp, db.max_timestamp"
+ ],
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "Ahk12agU3HLA",
+ "outputId": "6d93a289-6063-4cd6-fc66-72076c7a8bd1"
+ },
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "(Timestamp('1950-05-13 00:00:00'), Timestamp('2009-11-01 11:00:00'))"
+ ]
+ },
+ "metadata": {},
+ "execution_count": 6
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "1950 is the first season for F1! So we have data for the full history of F1. Note that the `max_timestamp` is the same as `test_timestamp`."
+ ],
+ "metadata": {
+ "id": "oKmC3R-H3OIU"
+ }
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "Next let's check out the tables in the database.\n",
+ "\n",
+ "More info on the schemas for F1 and all other datasets can be found at https://relbench.stanford.edu/."
+ ],
+ "metadata": {
+ "id": "zmGcKvxX3SO5"
+ }
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "We have the following tables:"
+ ],
+ "metadata": {
+ "id": "7RobvEoY3XQt"
+ }
+ },
+ {
+ "cell_type": "code",
+ "source": [
+ "db.table_dict.keys()"
+ ],
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "LyAdmIry3UFl",
+ "outputId": "dde952af-3953-47d5-8831-98676018cc65"
+ },
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "dict_keys(['constructors', 'results', 'standings', 'constructor_results', 'drivers', 'qualifying', 'races', 'circuits', 'constructor_standings'])"
+ ]
+ },
+ "metadata": {},
+ "execution_count": 7
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "That's 9 tables total! Let's look more closely at one of them."
+ ],
+ "metadata": {
+ "id": "3sdJ6OgukDrL"
+ }
+ },
+ {
+ "cell_type": "code",
+ "source": [
+ "table = db.table_dict[\"drivers\"]\n",
+ "table"
+ ],
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "n0OGtbMy3Y2n",
+ "outputId": "42f6ee62-f1ba-4519-bfec-32808baba1f3"
+ },
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "Table(df=\n",
+ " driverId driverRef code forename surname dob \\\n",
+ "0 0 hamilton HAM Lewis Hamilton 1985-01-07 \n",
+ "1 1 heidfeld HEI Nick Heidfeld 1977-05-10 \n",
+ "2 2 rosberg ROS Nico Rosberg 1985-06-27 \n",
+ "3 3 alonso ALO Fernando Alonso 1981-07-29 \n",
+ "4 4 kovalainen KOV Heikki Kovalainen 1981-10-19 \n",
+ ".. ... ... ... ... ... ... \n",
+ "852 852 mick_schumacher MSC Mick Schumacher 1999-03-22 \n",
+ "853 853 zhou ZHO Guanyu Zhou 1999-05-30 \n",
+ "854 854 de_vries DEV Nyck de Vries 1995-02-06 \n",
+ "855 855 piastri PIA Oscar Piastri 2001-04-06 \n",
+ "856 856 sargeant SAR Logan Sargeant 2000-12-31 \n",
+ "\n",
+ " nationality \n",
+ "0 British \n",
+ "1 German \n",
+ "2 German \n",
+ "3 Spanish \n",
+ "4 Finnish \n",
+ ".. ... \n",
+ "852 German \n",
+ "853 Chinese \n",
+ "854 Dutch \n",
+ "855 Australian \n",
+ "856 American \n",
+ "\n",
+ "[857 rows x 7 columns],\n",
+ " fkey_col_to_pkey_table={},\n",
+ " pkey_col=driverId,\n",
+ " time_col=None)"
+ ]
+ },
+ "metadata": {},
+ "execution_count": 8
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "The `drivers` table stores information on all F1 drivers that ever competed in a race. Note that the table comes with multiple bits of information:\n",
+ "- The table itself, `table.df` which is simply a Pandas DataFrame.\n",
+ "- The primary key column, `table.pkey_col`, which indicates that the `driverId` column holds the primary key for this particular table in the database.\n",
+ "- The primary time column, `table.time_col` which, if the entity is an event, records the time an event happened. In the case of drivers, they are non-temporal entities, so `table.time_col=None`.\n",
+ "- The other tables that foreign keys points to `table.fkey_col_to_pkey_table`. If the table has any foreign key columns, then this dict indicates which table we foreign key corresponds to. Again in the case of drivers this is not applicable.\n",
+ "\n",
+ "We can start to explore the data a little, e.g., check out the oldest and youngest ever F1 drivers, spanning 3 centuries!"
+ ],
+ "metadata": {
+ "id": "ZPEZo1TNkMaU"
+ }
+ },
+ {
+ "cell_type": "code",
+ "source": [
+ "table.df.iloc[table.df[\"dob\"].idxmax()]"
+ ],
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "paE62mcKkDCF",
+ "outputId": "f6535bd4-059f-4f87-bb60-b5ac05cc3889"
+ },
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "driverId 855\n",
+ "driverRef piastri\n",
+ "code PIA\n",
+ "forename Oscar\n",
+ "surname Piastri\n",
+ "dob 2001-04-06 00:00:00\n",
+ "nationality Australian\n",
+ "Name: 855, dtype: object"
+ ]
+ },
+ "metadata": {},
+ "execution_count": 9
+ }
+ ]
+ },
+ {
+ "cell_type": "code",
+ "source": [
+ "table.df.iloc[table.df[\"dob\"].idxmin()]"
+ ],
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/"
+ },
+ "id": "pqxywakH29js",
+ "outputId": "a5c2d066-3700-43da-83c1-0fd6828ceca8"
+ },
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ "driverId 741\n",
+ "driverRef etancelin\n",
+ "code \\N\n",
+ "forename Philippe\n",
+ "surname Étancelin\n",
+ "dob 1896-12-28 00:00:00\n",
+ "nationality French\n",
+ "Name: 741, dtype: object"
+ ]
+ },
+ "metadata": {},
+ "execution_count": 10
+ }
+ ]
+ },
+ {
+ "cell_type": "markdown",
+ "source": [
+ "Going back to the `table.time_col` and `table.fkey_col_to_pkey_table`, the `results` table contains a non-trivial example."
+ ],
+ "metadata": {
+ "id": "VJ3Se_Irkff1"
+ }
+ },
+ {
+ "cell_type": "code",
+ "source": [
+ "table = db.table_dict[\"results\"]\n",
+ "table.df"
+ ],
+ "metadata": {
+ "colab": {
+ "base_uri": "https://localhost:8080/",
+ "height": 791
+ },
+ "id": "DSmDQc3Pke5c",
+ "outputId": "726f3491-5436-4a95-a905-607279afd7b6"
+ },
+ "execution_count": null,
+ "outputs": [
+ {
+ "output_type": "execute_result",
+ "data": {
+ "text/plain": [
+ " resultId raceId driverId constructorId number grid position \\\n",
+ "0 0 0 660 152 18.0 21 11.0 \n",
+ "1 1 0 790 149 8.0 12 NaN \n",
+ "2 2 0 579 49 1.0 3 NaN \n",
+ "3 3 0 661 149 9.0 10 NaN \n",
+ "4 4 0 789 152 17.0 7 NaN \n",
+ "... ... ... ... ... ... ... ... \n",
+ "20318 20318 819 1 1 6.0 8 5.0 \n",
+ "20319 20319 819 21 22 23.0 4 4.0 \n",
+ "20320 20320 819 17 22 22.0 5 3.0 \n",
+ "20321 20321 819 16 8 14.0 3 2.0 \n",
+ "20322 20322 819 2 2 16.0 9 9.0 \n",
+ "\n",
+ " positionOrder points laps milliseconds fastestLap rank statusId \\\n",
+ "0 11 0.0 64 NaN NaN NaN 16 \n",
+ "1 21 0.0 2 NaN NaN NaN 126 \n",
+ "2 12 0.0 62 NaN NaN NaN 44 \n",
+ "3 20 0.0 5 NaN NaN NaN 6 \n",
+ "4 19 0.0 8 NaN NaN NaN 51 \n",
+ "... ... ... ... ... ... ... ... \n",
+ "20318 5 4.0 55 5669667.0 54.0 7.0 1 \n",
+ "20319 4 5.0 55 5666149.0 54.0 4.0 1 \n",
+ "20320 3 6.0 55 5661881.0 49.0 6.0 1 \n",
+ "20321 2 8.0 55 5661271.0 14.0 5.0 1 \n",
+ "20322 9 0.0 55 5689355.0 49.0 15.0 1 \n",
+ "\n",
+ " date \n",
+ "0 1950-05-13 00:00:00 \n",
+ "1 1950-05-13 00:00:00 \n",
+ "2 1950-05-13 00:00:00 \n",
+ "3 1950-05-13 00:00:00 \n",
+ "4 1950-05-13 00:00:00 \n",
+ "... ... \n",
+ "20318 2009-11-01 11:00:00 \n",
+ "20319 2009-11-01 11:00:00 \n",
+ "20320 2009-11-01 11:00:00 \n",
+ "20321 2009-11-01 11:00:00 \n",
+ "20322 2009-11-01 11:00:00 \n",
+ "\n",
+ "[20323 rows x 15 columns]"
+ ],
+ "text/html": [
+ "\n",
+ " | \n", + " | resultId | \n", + "raceId | \n", + "driverId | \n", + "constructorId | \n", + "number | \n", + "grid | \n", + "position | \n", + "positionOrder | \n", + "points | \n", + "laps | \n", + "milliseconds | \n", + "fastestLap | \n", + "rank | \n", + "statusId | \n", + "date | \n", + "
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | \n", + "0 | \n", + "0 | \n", + "660 | \n", + "152 | \n", + "18.0 | \n", + "21 | \n", + "11.0 | \n", + "11 | \n", + "0.0 | \n", + "64 | \n", + "NaN | \n", + "NaN | \n", + "NaN | \n", + "16 | \n", + "1950-05-13 00:00:00 | \n", + "
| 1 | \n", + "1 | \n", + "0 | \n", + "790 | \n", + "149 | \n", + "8.0 | \n", + "12 | \n", + "NaN | \n", + "21 | \n", + "0.0 | \n", + "2 | \n", + "NaN | \n", + "NaN | \n", + "NaN | \n", + "126 | \n", + "1950-05-13 00:00:00 | \n", + "
| 2 | \n", + "2 | \n", + "0 | \n", + "579 | \n", + "49 | \n", + "1.0 | \n", + "3 | \n", + "NaN | \n", + "12 | \n", + "0.0 | \n", + "62 | \n", + "NaN | \n", + "NaN | \n", + "NaN | \n", + "44 | \n", + "1950-05-13 00:00:00 | \n", + "
| 3 | \n", + "3 | \n", + "0 | \n", + "661 | \n", + "149 | \n", + "9.0 | \n", + "10 | \n", + "NaN | \n", + "20 | \n", + "0.0 | \n", + "5 | \n", + "NaN | \n", + "NaN | \n", + "NaN | \n", + "6 | \n", + "1950-05-13 00:00:00 | \n", + "
| 4 | \n", + "4 | \n", + "0 | \n", + "789 | \n", + "152 | \n", + "17.0 | \n", + "7 | \n", + "NaN | \n", + "19 | \n", + "0.0 | \n", + "8 | \n", + "NaN | \n", + "NaN | \n", + "NaN | \n", + "51 | \n", + "1950-05-13 00:00:00 | \n", + "
| ... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "
| 20318 | \n", + "20318 | \n", + "819 | \n", + "1 | \n", + "1 | \n", + "6.0 | \n", + "8 | \n", + "5.0 | \n", + "5 | \n", + "4.0 | \n", + "55 | \n", + "5669667.0 | \n", + "54.0 | \n", + "7.0 | \n", + "1 | \n", + "2009-11-01 11:00:00 | \n", + "
| 20319 | \n", + "20319 | \n", + "819 | \n", + "21 | \n", + "22 | \n", + "23.0 | \n", + "4 | \n", + "4.0 | \n", + "4 | \n", + "5.0 | \n", + "55 | \n", + "5666149.0 | \n", + "54.0 | \n", + "4.0 | \n", + "1 | \n", + "2009-11-01 11:00:00 | \n", + "
| 20320 | \n", + "20320 | \n", + "819 | \n", + "17 | \n", + "22 | \n", + "22.0 | \n", + "5 | \n", + "3.0 | \n", + "3 | \n", + "6.0 | \n", + "55 | \n", + "5661881.0 | \n", + "49.0 | \n", + "6.0 | \n", + "1 | \n", + "2009-11-01 11:00:00 | \n", + "
| 20321 | \n", + "20321 | \n", + "819 | \n", + "16 | \n", + "8 | \n", + "14.0 | \n", + "3 | \n", + "2.0 | \n", + "2 | \n", + "8.0 | \n", + "55 | \n", + "5661271.0 | \n", + "14.0 | \n", + "5.0 | \n", + "1 | \n", + "2009-11-01 11:00:00 | \n", + "
| 20322 | \n", + "20322 | \n", + "819 | \n", + "2 | \n", + "2 | \n", + "16.0 | \n", + "9 | \n", + "9.0 | \n", + "9 | \n", + "0.0 | \n", + "55 | \n", + "5689355.0 | \n", + "49.0 | \n", + "15.0 | \n", + "1 | \n", + "2009-11-01 11:00:00 | \n", + "
20323 rows × 15 columns
\n", + "