{ "cells": [ { "cell_type": "markdown", "id": "65f79301-6b74-4df9-91a3-d8ab060bbe9e", "metadata": {}, "source": [ "# TEEHR Evaluation Example 2 - Continued\n", "## Daily Data, NWM 3.0 Retrospective and MARRMoT_37 HBV, CAMELS Subset (542)\n", "\n", "### Evaluate Model Output\n", "The prior notebooks were all about setting the stage for evaluation - preparing and joining the data to make the evaluation as easy and efficient as possible. The reality is that evaluation is not remotely easy, particularly when you are dealing with very large datasets, 1000s of locations, many different models to compare, different baselines, different types of variables, different decision objectives, etc. There are a nearly endless number of ways to slice up the data, calculate metrics and visualize comparisons. It is difficult (or impossible) to know in advance which approach is going to give us the most useful insights to answer a given question (e.g., which model is better for certain conditions and objectives? how does it compare to a particular baseline? is there a relationship between performance and location characteristics (attributes))?\n", "\n", "In this notebook we will demonstrate how to use TEEHR to calculate metrics from the joined TEEHR database created in Notebook ex2-1, using a range of different options for grouping and filtering. We will then create some common graphics based on the results.\n", "\n", "\n", "#### In this notebook we will perform the following steps:\n", "