Avara starts with a county-scoped parcel record
Avara does not begin equity analysis with a disconnected spreadsheet or an unexplained model output. The platform records county data sources as ingestion runs and snapshots, then turns raw parcel artifacts into canonical and enriched parcel indexes used by search, GIS, analytics, and model workflows. That creates one county-scoped foundation for the values and evidence shown in Vantage.
The distinction matters. A reviewer needs to know whether a surprising result came from the valuation method, a stale assessment, a missing characteristic, a recent sale, permit activity, or a source-record conflict. By keeping source identity and pipeline context attached to the parcel, Avara gives reviewers a way to investigate the record before treating the output as a conclusion.
Equalization protects the validation evidence
Avara’s production Equalization workflow separates model fitting, residual calibration, and validation into chronological, non-overlapping windows. Validation sales are not used for training, time adjustment, outlier handling, early stopping, or candidate selection. The time trend is learned only from the fit window, while preprocessing statistics and calibration rules are retained with the model bundle.
The workflow also applies factor clamps and keeps calibration group-aware instead of allowing individual sold parcels to overwrite the method. For historical predictive backtesting, Avara shifts the full methodology to a prior valuation year, freezes that roll, and evaluates it only against later qualified sales. Those controls make the evidence more useful because a strong score cannot be created by leaking the answer into the model.
Equity Monitor keeps the comparison fair
Inside Vantage, Equity Monitor compares assessed totals and equalized values against the same qualified sale anchors. Teams can review Assessment-to-Sale Ratio, Coefficient of Dispersion, and Price-Related Differential before and after equalization without changing the underlying comparison sample between views.
The monitor supports countywide and segmented review across study groups, ZIP codes, zoning, and other available county slices. Each segment carries the supporting sample size, county share, record completeness, and residual-risk context. Avara labels the evidence as qualified, use caution, or thin sample so a numerically attractive result is not presented with more certainty than its data can support.
- ASR shows the relationship between assessed or equalized values and qualified sales.
- COD shows dispersion within the selected county segment.
- PRD helps reviewers examine vertical equity across value levels.
- CSV and JSON exports preserve the selected segment context for follow-up review.
Avara separates equity auditing from value production
Avara’s assessment-equity audit is intentionally separate from the model that generates equalized values. Socioeconomic and protected-class context can be used to study historical assessment gaps, but it is not fed back into the valuation workflow. The audit estimates and explains the difference between prior assessed values and the equalized baseline, with margins of error and data limitations treated as part of the analysis.
This separation supports a clearer public-sector standard: valuation inputs remain valuation inputs, while equity analysis tests the historical pattern around the result. The output is statistical evidence for qualified review, not a claim of causation and not an undisclosed adjustment to an individual parcel.
The result becomes an assessor review workflow
Vantage turns equity findings into parcel-level work. Its review-priority signals can combine valuation residuals, assessed-to-equalized changes, stale assessments, later permit activity, recent-sale gaps, large value deltas, and missing or low-confidence data. Outliers adds peer, transaction, ownership, permit, property-change, and assessment-history evidence without presenting a review candidate as a final determination.
From there, Parcel Audit follows the value from the source assessment through the model factor, sales calibration, ratio checks, and final equalized value. Reviewers can inspect source fields, peer context, value-driver effects, comparable evidence, model/run provenance, and the source hash. Equity is therefore not a dashboard detached from operations; it is connected to the exact parcels and evidence that staff can review.