Data type
| Option | Description |
|---|---|
Spatial only (cross-sectional) | Uses a single time point - or a collapsed period - across spatial units. |
Time-series (one sample) | Analyzes repeated measurements over time for a single unit. |
Panel data (many samples) | Combines cross‑sectional and time‑series variation across multiple units. |
Definition
The data structure(s) that the method can appropriately handle, such as spatial cross‑sectional measurements, single‑unit time‑series, or multi‑unit panel data.
Explanation
Matching the method to your data’s form is crucial: some approaches require only a snapshot in space, others demand temporal repetition for a single unit, and still others exploit variation across many sampled units over time. Choosing an incompatible method can invalidate results or leave valuable structure unused.
Tools/rationale for helping assessment
- Draw a causal graph to make sure no relevant to the system and easily available variable has not been forgotten.
- Inspect your data: Do you have only one timestamp per pixel (cross-sectional), a single‐site time‐series, or many sites over time?
- Use visualization helpers like a covariate alignment graph to ensure good comprehension of the data structure.
- Prototype eventually the chosen method from the suggested set on a small subset of your data to confirm compatibility.
Example
A researcher has 10 years’ annual species richness from 30 protected areas - that’s panel data - so they exclude methods dealing only with a single‐site time-series or a single time point.