FIT5137 Chap.10 Spatial Visualisation, Indexes and Catchments
Spatial Visualisation, Indexes and Catchments
Spatial Visualisation, Indexes and Catchments combines data loading, QGIS inspection, advanced query decomposition, indexing and cautious interpretation. A catchment can mean straight-line proximity, network reachability, administrative assignment or observed origins, so the rule, unit and boundary policy must be declared.
Reliable workflows expose intermediate populations after loading, validity checks, transformation, candidate selection and exact predicates. Spatial indexes reduce candidate work through bounded search, but the exact predicate still decides membership. Optimisation is accepted only when qualifying identifiers remain unchanged and representative query plans and timings support the improvement.
Visualisation closes the evidence chain rather than replacing it: symbols, classification, projection, extent, legend and coverage shape interpretation. Crowdsource records require explicit limits because no reports can mean low hazard incidence or low participation. The chapter ends by making maps reproducible through stored query definitions, coordinate settings, classification breaks and extraction context.
What this chapter covers
- 01
Define the catchment rule, unit and boundary policy
- 02
Reconcile populations at every query stage
- 03
Preserve stable identifiers through transformation
- 04
Separate candidate pruning from exact predicates
- 05
Inspect plans, selectivity and repeated timings
- 06
Keep logical results unchanged during optimisation
- 07
Match visual variables to data meaning
- 08
Qualify maps by coverage, scope and uncertainty
Worked example · free
Audit an overlapping catchment result
- 2Total assignments equal 18 plus 24 plus 21 plus 17, which is 80.
- 2Each of the 10 overlapping points is counted one extra time in the assignment total because it appears under two hubs.
- 1Unique-point count equals 80 minus 10 duplicate assignments, which is 70.
- 1Check that 60 points have one assignment and 10 points have two assignments: 60 plus 10 equals 70 unique points.
- 1Recompute assignments from that distribution: 60 times 1 plus 10 times 2 equals 80.
- 1Report both 80 assignments and 70 unique points because they answer different questions about coverage and overlap.
Key terms
- Catchment rule
- The declared spatial logic that assigns locations to a facility or service area.
- Candidate set
- The reduced group of geometry pairs sent to an exact predicate for final evaluation.
- Query plan
- The database execution strategy used to access, join, filter and aggregate rows.
- Spatial selectivity
- The degree to which a spatial filter reduces candidates relative to the full population.
- Choropleth map
- A thematic map shading regions by a normalised or otherwise meaningful area value.
- Coverage bias
- A distortion caused by uneven observation or contribution rather than the phenomenon itself.
Spatial Visualisation, Indexes and Catchments FAQ
What does a spatial index change?
It changes how the database finds likely candidates, often using bounding regions. It must not change the logical definition of the result. The exact relationship or distance predicate still determines qualification, and identifier sets should match before and after tuning.
Why can a catchment count exceed unique locations?
Catchments can overlap, so one location may be assigned to several hubs. Assignment count measures hub-location relationships, while a distinct count measures locations. Report both or apply a transparent single-assignment policy if the decision requires exclusivity.
Is a map enough to validate a query?
No. Maps reveal displacement, gaps and outliers but can hide coincident points, tiny boundary errors and classification choices. Pair visual inspection with stage counts, stable identifiers, known cases, SQL reconciliation and reference-system checks.
How should crowdsource gaps be interpreted?
No report does not prove no event. Participation, observation opportunity and reporting habits may vary across space. Compare contributor activity or coverage, disclose the limitation and seek independent evidence before making a claim about underlying incidence.
Assessment move
Rehearse catchment work as a staged ledger: input, valid geometry, transformed geometry, candidate pairs, exact matches, distinct identifiers and grouped output. Write the count at every stage and preserve sample exclusions. Learn indexing conceptually as candidate pruning followed by exact evaluation, then inspect query plans rather than promising speed from index presence alone.
Rebuild one map from saved query, coordinate system and classification breaks. For every visual conclusion, write an alternative explanation based on coverage or symbology and name the extra evidence that could distinguish them. Add overlapping catchments and compute both assignment and unique-feature counts. If exclusive allocation is required, introduce a tie and document the secondary rule.
Compare plans over a small extent and a broad extent, recording candidates, exact matches and repeated timings while verifying identical result keys. Finally, redesign one misleading map by changing raw counts to an appropriate rate, clarifying the legend and adding a coverage layer. Explain which visual inference changed and which database result remained constant.
Save the extraction time, query, coordinate system, class breaks and legend wording as a reproducibility bundle. Recreate the visual from that bundle after changing one parameter so the effect of method is distinguishable from the effect of new data. Record both unique-feature and assignment counts when catchments overlap so the visual legend cannot conceal duplicated membership.
