{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Handling dimensions with scores\n",
"\n",
"`Scores` is primarily designed to work with xarray data. xarray allows us to work with multidimensional data with just a single implementation of a metric. In this notebook we illustrate how scores can work with data with different dimensions.\n",
"\n",
"We recommend that you visit the [xarray documentation](https://docs.xarray.dev/en/stable/getting-started-guide/quick-overview.html) to learn about xarray if you aren't familiar with xarray.\n",
"\n",
"First, let's illustrate how scores can be used with data with different dimensions using synthetic data."
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"import xarray as xr\n",
"import numpy as np\n",
"import pandas as pd\n",
"\n",
"from scores.continuous import mse\n",
"\n",
"np.random.seed(100)"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"# First, let's set up some coordinates\n",
"valid_times = pd.date_range(\"2024-01-01\", \"2024-01-31\")\n",
"lead_time = np.arange(1, 11)\n",
"station_numbers = np.arange(1, 11)\n",
"lat = np.arange(-90, 90, 15)\n",
"lon = np.arange(-180, 180, 15)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Data could be 1-dimensional (e.g., a timeseries at an airport)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
"
],
"text/plain": [
"\n",
"array(1.07490488)"
]
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"mse(fcst_stations, obs_stations, reduce_dims=\"all\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"If we want to calculate the MSE aggregated across space (`station_number` dimension), but not time (`valid_time` dimension), then there are also two options.\n",
"\n",
"1. Use `preserve_dims=[\"valid_time\"]`, or \n",
"2. Use `reduce_dims=[\"station_number']` \n",
"\n",
"when the score is called.\n",
"\n",
"You cannot supply both the `reduce_dims` and `preserve_dims` args at the same time."
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"