https://mtaniguchiking.github.io/M4MD-forecast-docs-dev/docs/5-forecasting/running/guides/Recent content in Guides on Models for missing dataHugoen-ushttps://mtaniguchiking.github.io/M4MD-forecast-docs-dev/docs/5-forecasting/running/guides/mock-data/Mon, 01 Jan 0001 00:00:00 +0000https://mtaniguchiking.github.io/M4MD-forecast-docs-dev/docs/5-forecasting/running/guides/mock-data/<h1 id="your-first-forecast-mock-data"> Your First Forecast (Mock Data) <a class="anchor" href="#your-first-forecast-mock-data">#</a> </h1> <p>This is the simpliest path to perform a forecast. All data is pre-committed, so there&rsquo;s nothing to download or prepare. The walkthrough has three steps.</p> <h2 id="prerequisites"> Prerequisites <a class="anchor" href="#prerequisites">#</a> </h2> <p>A working M4MD repo. [TODO: formalize this]</p> <h2 id="step-0-understand-the-mock-data"> Step 0: Understand the mock data <a class="anchor" href="#step-0-understand-the-mock-data">#</a> </h2> <p>The mock dataset represents a single park unit (<strong>ELDO</strong>) with two vegetation strata (<strong>A</strong> and <strong>B</strong>), three sites per stratum, and three transects per site — surveyed annually from <strong>2000 to 2025</strong>. Each transect records how many of 100 sample points hit a plant (<code>y_hits</code> out of <code>n_points = 100</code>). A single climate covariate, <strong>precipitation (ppt)</strong>, declines gradually over the training period and drives the vegetation response.</p>https://mtaniguchiking.github.io/M4MD-forecast-docs-dev/docs/5-forecasting/running/guides/climate-futures-data/Mon, 01 Jan 0001 00:00:00 +0000https://mtaniguchiking.github.io/M4MD-forecast-docs-dev/docs/5-forecasting/running/guides/climate-futures-data/<h1 id="forecasting-with-real-world-data"> Forecasting with Real-World Data <a class="anchor" href="#forecasting-with-real-world-data">#</a> </h1> <!-- TODO: Intro paragraph — this path is for users who have their own NPS monitoring data and want to run a real forecast. It has a moderate lift: the user provides a response variable CSV, and an external notebooks repo handles covariate and scenario data preparation. --> <h2 id="what-youll-need"> What You&rsquo;ll Need <a class="anchor" href="#what-youll-need">#</a> </h2> <!-- TODO: Checklist — same R/JAGS prereqs as mock walkthrough, plus: - your own response variable CSV (format TBD — link to column schema) - access to the external notebooks repo (link TBD) --> <h2 id="step-1-prepare-your-covariate-and-scenario-data"> Step 1: Prepare Your Covariate and Scenario Data <a class="anchor" href="#step-1-prepare-your-covariate-and-scenario-data">#</a> </h2> <!-- TODO: Point to the external notebooks repo (URL TBD). Explain: - User provides their response variable CSV - The notebooks produce a formatted covariates CSV and a scenarios CSV - The output format matches what analysis-pipeline.R and forecast-pipeline.R expect --> <h2 id="step-2-fit-the-model"> Step 2: Fit the Model <a class="anchor" href="#step-2-fit-the-model">#</a> </h2> <p>Once your data files are ready, follow the same fit step as the mock walkthrough:</p>https://mtaniguchiking.github.io/M4MD-forecast-docs-dev/docs/5-forecasting/running/guides/custom-data/Mon, 01 Jan 0001 00:00:00 +0000https://mtaniguchiking.github.io/M4MD-forecast-docs-dev/docs/5-forecasting/running/guides/custom-data/<h1 id="running-a-forecast-with-custom-data"> Running a Forecast with Custom Data <a class="anchor" href="#running-a-forecast-with-custom-data">#</a> </h1> <!-- TODO: Intro paragraph — this is the full reference for users bringing entirely their own data (response, covariates, and scenarios). No external tool required. This page is denser than the walkthroughs and serves as a reference rather than a tutorial. --> <h2 id="required-input-files"> Required Input Files <a class="anchor" href="#required-input-files">#</a> </h2> <!-- TODO: Describe each input file with column schemas: 1. Response variable CSV — columns, units, required fields 2. Covariates CSV (training) — columns, format 3. Scenarios CSV (future) — columns, scenario_id vs model_run_id distinction, how multiple GCM runs are represented Cross-reference forecasting/getting-started/generate-mock-data.R for concrete examples of each format. --> <h2 id="forecast-config-yaml-reference"> Forecast Config YAML Reference <a class="anchor" href="#forecast-config-yaml-reference">#</a> </h2> <!-- TODO: Document every key in the forecast config YAML: - Required keys (paths to inputs, output dir, likelihood) - Covariate strategy options: provided / hold_last / hold_mean - Scenario filtering options - n_cores, seed, etc. Source: forecasting/forecast/read-forecast-config.R --> <h2 id="running-the-pipeline"> Running the Pipeline <a class="anchor" href="#running-the-pipeline">#</a> </h2> <div class="highlight"><pre tabindex="0" style="color:#f8f8f2;background-color:#272822;-moz-tab-size:4;-o-tab-size:4;tab-size:4;-webkit-text-size-adjust:none;"><code class="language-bash" data-lang="bash"><span style="display:flex;"><span>Rscript forecasting/forecast/forecast-pipeline.R <span style="color:#ae81ff">\ </span></span></span><span style="display:flex;"><span> --config path/to/forecast-config.yaml </span></span></code></pre></div><!-- TODO: Note any useful flags (--dry-run, --verbose, etc. if they exist). Describe what a successful run looks like in the console. --> <h2 id="output-files"> Output Files <a class="anchor" href="#output-files">#</a> </h2> <!-- TODO: Document the output directory layout and key files: - Trajectory plots (per-site, per-stratum) - Spaghetti plots - forecast-site-summaries.csv — describe each column - mu_median vs y_rep_median — explain the distinction in plain English Source: forecasting/forecast/forecast-plot-engine.R -->