How We Build Exposure Profiles

What This Tool Does

The Community Resilience Index builds an exposure profile for every Australian postcode. It answers three questions:

1. What is your community’s structural shape?The characteristics that determine how pressures reach you: car dependency, distance from fuel supply, industry concentration, remoteness, housing affordability, energy independence.
2. Which pressures hit your community hardest?Algorithmic exposure mapping across six domains: fuel, food, electricity, economic, housing, emergency. Driven by structural data, not opinion.
3. What should you do and what should you watch?Actions ranked by urgency and live signals contextualised to your specific community.

We deliberately do not produce a single composite score. A number hides the most useful information: why your community is exposed and what you can do about it. A profile shows the shape of your exposure. Shape is actionable; a number is not.

Three Layers, Not One Score

Layer 1: Structural Profile

We extract up to 11 characteristics per postcode from official Australian data:

• Car dependency: proportion of commuters who drive (ABS Census 2021)
• Industry diversity: Shannon index across employment sectors (ABS Census 2021)
• Agricultural workforce: proportion in agriculture, forestry, fishing (ABS Census 2021)
• Remoteness: Modified Monash Model category (ABS 2023)
• Housing stress: proportion of households spending >30% of income on housing (ABS Census 2021)
• Solar capacity: installed kW per postcode (Clean Energy Regulator 2024)
• Socioeconomic index (IRSD): ABS SEIFA 2021
• Median household income: weekly (ABS Census 2021)
• Transport mode diversity: Shannon index across commute modes (ABS Census 2021)
• Internet access: proportion of dwellings connected (ABS Census 2021)

Each characteristic is compared against all Australian postcodes so you can see where your community sits nationally. The 2–3 characteristics that deviate most from the national median are surfaced as your community’s most distinctive features.

Layer 2: Exposure Mapping

Structural characteristics are mapped to exposure domains using transparent, algorithmic rules grounded in published research from the ACCC, BITRE, ABARES, and the RBA. These are not machine-learned or LLM-generated. Every rule is documented and auditable. Full evidence base on the references page.

Each domain receives an exposure weight between 0 and 1. Domains are ranked so you see your highest exposures first. Every weight includes a plain-language explanation of what’s driving it.

Layer 3: Contextualised Signals

Live data feeds are ranked by relevance to your community’s structural profile. The same signal means different things in different places:

• A crude oil price spike matters more in a regional postcode with 80% car dependency than in an inner-city area with strong public transport.
• Farm input cost increases hit harder in communities where 15% of workers are in agriculture.
• RBA rate changes compound most in postcodes with high housing stress.

Signal context is generated using parameterised templates. deterministic, auditable text that fills in your community’s actual values. No generative AI is used.

How Actions Are Generated

Actions are computed from exposure weights, structural drivers, and diversity analysis. The formula:

action_urgency = base_score × exposure_weight × (1 + entrainment_penalty)

Where base_score reflects how critical the action type is inherently, exposure_weight is how exposed your specific postcode is in that domain, and entrainment_penalty increases urgency for communities with concentrated dependencies (0.0 for diversified, 0.15 for moderate, 0.3 for concentrated).

Actions are categorised as Household (things you can do alone), Community (things that need collective action), or Advocacy (things that need structural change). They are time-sequenced: Do now, This month, Ongoing.

Concentrated or Diversified: Why This Matters

This is where our approach differs from standard resilience tools. Most indices treat all forms of community strength as equal. A postcode where 90% of workers are in one industry scores the same on “employment rate” as one where workers are spread across many industries. Both look resilient by that measure.

But they fail differently. The single-industry community is concentrated: locked into one dependency. It looks stable right up until that industry contracts, at which point everything falls apart at once. There is nothing else to reorganise around.

The diverse-industry community is diversified: its parts are connected but not locked together. When one sector struggles, others can absorb the shock. People have options. The community can reorganise.

We measure diversity using the Shannon diversity index across industry employment and transport modes. Higher diversity means more ways for a community to reorganise under stress. This is presented as a spectrum, not a binary.

Diversity directly affects the urgency of recommended actions. Communities with concentrated dependencies face a higher entrainment penalty in the action ranking formula, because when their dominant system fails there is no fallback.

Cascade Timeline Estimates

When global prices change, communities don’t feel it immediately. Crude oil takes weeks to reach the bowser. Farm input costs take months to flow through to food prices. Rate changes take weeks to hit mortgage repayments.

We provide estimated propagation timelines for each domain, adjusted for remoteness. These are based on supply chain structure, ACCC fuel monitoring data, AEMO market analysis, ABARES commodity reports, and RBA transmission research. They are estimates and projections, not precise forecasts.

Data Sources and Vintage

Most structural data comes from the 2021 Census. This is a five-year-old snapshot. Communities change. The index measures structural conditions as of that collection point. The 2026 Census (August) will provide a significant refresh.

Non-census sources (solar capacity, remoteness classification) are more recent (2023–2024). Live signals are fetched in real-time from public APIs.

Every data point on the profile page includes its source and vintage. We show what data is available and name what is missing because a partial picture honestly labelled is more useful than a complete-looking picture hiding its gaps.

Limitations

• Most structural data is from the 2021 Census, now five years old. Communities change. Treat structural characteristics as a baseline, not a current snapshot.
• Exposure mapping uses algorithmic rules that may not capture every local factor. A community next to a military base or major logistics hub may have supply chain resilience not reflected in the census data.
• Cascade timeline estimates are based on industry structure and historical patterns, not real-time supply chain monitoring. Actual propagation depends on market conditions, contracts, and policy interventions.
• Signal data depends on public API availability. Some feeds (notably state fuel price APIs) are intermittently unreliable. We show the last successful fetch time.
• Actions are computed from structural data and exposure rules. They are starting points for community conversation, not prescriptions. Local knowledge matters more than any index.