---
name: brycewang-stanford/agsy-figures-and-tables
source: https://app.decimal.ai/s/brycewang-stanford-agsy-figures-and-tables@1/SKILL.md
source_sha256: a7bfe476f976
---

# Figures & Tables (agsy-figures-and-tables)

Exhibits are where a systems reviewer checks whether the model and the trade-offs are real. At AgSy the
strongest exhibits make **interactions, dynamics, and trade-offs** visible, and let a reader judge
**model performance** at a glance. A wall of single-treatment bar charts does not belong in a systems
paper.

## When to trigger

- Designing the main results figure or a key model-performance exhibit
- Deciding what belongs in the article vs. supplementary material
- A reviewer found an exhibit unclear, mislabeled, or not self-contained
- Presenting scenario comparisons or trade-offs

## Principles

1. **Self-contained.** A reader should understand each exhibit from its caption, axis/column labels, and
   note alone. State units, system level (field/farm/landscape/region), scenario, N, and time horizon.
2. **Show the system, not one factor.** Prefer exhibits that reveal interactions and trade-offs:
   - **Trade-off / Pareto-front** plots (e.g., profit vs. N leaching across options)
   - **Observed-vs-simulated (1:1)** plots with fit statistics (RMSE/NSE/bias) for model evaluation
   - **Time series** of state variables to show dynamics and feedbacks
   - **Sankey / flow** diagrams for resource, nutrient, or energy flows
   - **Conceptual system diagrams** (boxes/arrows) for boundaries and components
   - **Multi-indicator / radar** profiles for sustainability comparisons
3. **Show uncertainty.** Intervals, ensemble ranges, or boxplots — not bare point estimates from a
   single model run.
4. **Accessible.** Colorblind-safe palettes; legible in grayscale; no chartjunk or 3D; vector output
   (PDF/EPS) for print.
5. **Reproducible.** Each exhibit is generated by the master workflow; numbers match the deposited
   data/code/model exactly (see `agsy-reproducibility-and-data-policy`).

## Article vs. supplementary
- Keep the few exhibits that carry the systems argument (trade-offs, key dynamics, model fit) in the
  article; move full calibration tables, parameter lists, and exhaustive scenario grids to
  **supplementary material**.

## Pick the exhibit from the claim (illustrative)

A crop–livestock simulation has four candidate main figures; choose by what the systems claim needs:

- *Claim is "the model is credible"* → observed-vs-simulated 1:1 plot with RMSE/NSE/bias on the figure.
- *Claim is "there is a trade-off"* → Pareto front of gross margin vs. N surplus across options, with
  the recommended region marked — not a scenario bar chart that hides the frontier.
- *Claim is "feedbacks drive the result"* → time series of state variables (herd, manure N, rotation)
  showing the loop, with an ensemble band for uncertainty.
- *Claim is "this is a system"* → a conceptual boundary-and-components diagram, often the graphical
  abstract.

## Referee pushback → the AgSy-specific fix

- *"The main figure is a single-treatment bar chart."* → Replace it with a trade-off or 1:1 exhibit that
  shows interactions, not one factor.
- *"Fit is asserted, not shown."* → Pair the observed-vs-simulated plot with the fit statistics on the
  same figure.
- *"No uncertainty on the headline result."* → Add intervals, an ensemble band, or boxplots; one model
  run is not a result.

## Anti-patterns

- A single-treatment bar chart presented as the main result of a "systems" paper
- Observed-vs-simulated shown without fit statistics, or fit statistics with no plot
- Scenario tables that bury the trade-off instead of plotting the frontier
- Point estimates from one model run with no uncertainty shown
- Color-only encoding that fails in grayscale or for colorblind readers
- Exhibit numbers/values that don't match the deposited model output


## Operating pass for Agricultural Systems

Run this as a concrete capability pass. First lock the system boundary, actor decision, model/data linkage, and sustainability or food-security tradeoff; then test whether the manuscript addresses agricultural-systems reviewers who expect crop, farm, value-chain, environment, and policy components to be connected rather than listed.

- **Primary move:** Return a claim-evidence-risk ledger; every recommendation must point to a manuscript location or missing artifact.
- **Decision ledger:** return `claim / evidence / blocker / next edit` rows so the next pass can patch the manuscript directly.
- **Sibling comparison:** compare against Field Crops Research for plot-level agronomy, Global Food Security for policy synthesis, Agricultural Economics for economics-first work; if the neighboring outlet has the stronger audience claim, recommend re-routing before polishing.
- **Verification floor:** before submission-ready advice, re-open `resources/official-source-map.md` for volatile rules and name the one unresolved fact that could change the recommendation.

## Output format

```
【Main exhibit】what system interaction / trade-off / fit it shows
【Type】trade-off frontier / 1:1 / time series / flow diagram / system diagram / radar
【Self-contained?】caption + labels + units + level + scenario + N? [Y/N]
【Uncertainty shown?】intervals / ensemble / boxplot? [Y/N]
【Article vs supplementary】split decided
【Reproducible?】generated by master workflow, matches package? [Y/N]
【Next】agsy-writing-style
```

## Supplementary resources

- [`../../resources/external_tools.md`](../../resources/external_tools.md) — plotting, flow-diagram, and mapping tools for systems exhibits
- [`../../resources/official-source-map.md`](../../resources/official-source-map.md) — Highlights / graphical-abstract requirements