Accept, Review or Reject?

Chapter 1 Practicum — Accept, Review or Reject?

Professional outcome

By the end of this practicum, you will be able to make a traceable geospatial data-acceptance decision from incomplete evidence. You will deliver DATA_ACCEPTANCE_DECISION.md, the first formal gate in your UAV and Satellite Analysis Pipeline.

This practicum integrates Lessons 2.1–2.4. It is not a new software tutorial. Work as the scientist responsible for deciding what may enter an analysis.

1. The handover

The Baltic coastal-meadow team receives six assets:

Asset	Evidence received	Deliberate problem
A — management zones	polygon GeoPackage, EPSG:3301, producer note	accuracy not independently checked
B — UAV reflectance	referenced raster dimensions, transform and CRS	band names, units and calibration absent
C — plot candidates	CSV with x and y columns	CRS, axes and units absent
D — habitat map	5 cm output pixels	control-point review indicates poor positional accuracy
E — satellite scene	product metadata and valid CRS	acquisition is six weeks later than field sampling
F — legacy boundary	Shapefile components	incomplete provenance and truncated field descriptions

All files open successfully. That establishes only that software can parse them.

2. Use one decision rule

For a clearly stated intended use, assign one status:

• Accept for stated use: evidence supports the proposed operation, with limitations recorded
• Review before use: missing evidence could change the result and has a feasible owner or verification action
• Reject for stated use: a blocking risk cannot be resolved within the project, or the asset cannot support the proposed inference

A status applies to a use, not to a file forever. A layer rejected for 5 cm plot extraction might still be acceptable for a regional orientation map.

Check your understandingA raster has a CRS and opens correctly, but its acquisition date and band meaning are absent. What is the defensible decision?

Choose one answer

Review before scientific useAccept without conditionsReject and delete the source

Check answer

3. Build the evidence matrix

Create one row per asset with these columns:

Required field	Question
spatial model	What does one row, feature or cell represent?
CRS	Are reference system, axes and units verified?
support	What physical space contributes to one value?
accuracy and precision	What evidence describes closeness and repeatability?
resolution	What detail or sampling interval is represented?
temporal support	When, and over what interval, was the observation made?
provenance	Who produced it, from what source and process?
format	Does the format preserve the required structure and metadata?
remaining uncertainty	What claim remains unsupported?
decision	Accept, review or reject for the stated use?
next action and owner	What happens next, and who can resolve it?

Do not turn absence into a negative fact. “No accuracy statement supplied” is evidence; “the layer is inaccurate” is a conclusion that requires evidence.

Check your understandingWhat should an acceptance decision connect?

Choose one answer

Evidence, intended use, risk and next actionFilename and visual appearanceFile size and creation date only

Check answer

4. Predict before deciding

Before completing the matrix, predict which two assets are most likely to be rejected for plot-level field–imagery comparison. Then test that prediction against every evidence field. If your decision changes, document why. Changing a prediction in response to better evidence is good scientific practice.

5. Independent decision

Write DATA_ACCEPTANCE_DECISION.md with:

1. the intended analysis and its required spatial and temporal support
2. a six-row decision table
3. the blocking evidence for every review or rejection
4. one named owner and next action for each review
5. one short paragraph explaining why “opens successfully” is not an acceptance criterion

Keep the narrative to 250–350 words maximum, excluding the table. A concise decision is more useful than repeated definitions.

Check your understandingWhen is rejection appropriate?

Choose one answer

When a blocking risk cannot be resolved for the intended useWhenever metadata need reviewWhenever the format is unfamiliar

Check answer

Scientific interpretation

Acceptance means the available evidence is adequate for one declared use. It does not certify universal accuracy or ecological truth. Review means uncertainty is material but potentially resolvable. Rejection protects the analysis from a known blocking risk; it is not a punishment for imperfect data.

Reflection

1. Which asset looked strongest before you separated resolution from accuracy?
2. Which missing field most directly blocks field–image comparison?
3. How did temporal support alter a technically valid spatial decision?
4. What new evidence could change one rejection into acceptance?

Submission

• Decision record: DATA_ACCEPTANCE_DECISION.md
• Evidence: the completed matrix with sources and access dates
• Screenshot: one view showing decisions, risks and owners
• Written answer: the required 250–350-word decision narrative

Portfolio artifact

Artifact 2.A — Geospatial data acceptance decision

Place this decision gate before all processing stages in the UAV and Satellite Analysis Pipeline. Future assets should pass through the same evidence structure.