Workflow Narrative

Yield Data Conditioning and QA

Problem It Solves

How do we turn raw, noisy harvest monitor points into defensible, analysis-ready yield products for management decisions?

Who It Is For

Precision-ag service providers, agronomy analytics teams, and farm data engineering groups.

Primary User

Enterprise farms, precision agriculture consultancies, and digital agronomy platforms.

What It Does

Runs an end-to-end conditioning pipeline for noisy combine-harvester yield points.
Orchestrates edge QA, pass reconstruction, optional multi-header reconciliation, filtering, normalization, and swath map generation.
Produces contract-ready summary outputs for auditability and downstream zoning workflows.

How It Works

Flags/removes likely edge points using local neighborhood support.
Optionally filters telemetry anomalies using speed and heading consistency thresholds.
Auto-resolves common monitor-export field aliases when requested names are missing.
Reconstructs pass structure from point geometry and heading continuity, then smooths yield using pass-aware neighborhood statistics.
Optionally applies distance-based lag correction along reconstructed passes to compensate harvest monitor delay.
Optionally converts raw yield to a target moisture basis using moisture_field_name and target_moisture_pct.
Optionally applies robust MAD-based clipping on filtered yield values before normalization.
Propagates a QA confidence score to final point outputs, then builds swath polygons for map-ready products.
Indicative formula: filtered_yield ~= weighted_neighborhood(yield) with outlier replacement when z = |y - mean_adjacent_pass| / sd_adjacent_pass exceeds threshold.

Why It Wins

Provides a reproducible and transparent preprocessing chain with explicit QA outputs rather than opaque one-click cleaning.

Typical Buying Trigger

A team needs to standardize inconsistent yield data cleaning before zone generation, variable-rate planning, or year-over-year analytics.

Typical Presets

fast: lighter cleanup with larger tolerances for rapid turnaround.
balanced: default production profile.
strict: stronger edge/outlier controls for high-confidence analytics.

Inputs

Parameter	Optional	Description
input (yield point vector)	no	Raw yield telemetry points used as input to the conditioning and QA pipeline.
yield_field_name, moisture_field_name, target_moisture_pct, header_field_name	no	Field-name mappings for yield/moisture/header attributes used by cleaning stages.
use_field_aliases	no	Whether known attribute aliases should be resolved automatically.
speed_field_name, heading_field_name	no	Telemetry field mappings used for speed and heading quality checks.
min_speed_kmh, max_speed_kmh, max_heading_change_deg	no	Operating-speed and heading-change limits used for telemetry QC filtering.
profile: fast \| balanced \| strict	no	Operational profile controlling sensitivity and QA strictness for risk workflows.
swath_width, edge_radius, reconcile_radius, normalization_radius	no	Geometric parameters controlling swath-edge handling and neighborhood reconciliation.
lag_correction_mode: none \| distance	no	Lag correction mode controlling whether and how harvest lag compensation is applied.
lag_distance_m	no	Distance offset used when distance-based lag correction is enabled.
filtering_mode: standard \| robust	no	Outlier-filtering method selection for yield cleaning.
robust_mad_threshold, z_score_threshold, min_yield, max_yield	no	Statistical thresholds and hard limits used for yield outlier rejection.
optional mean_tonnage	yes	Optional mean tonnage override used during normalization/reconciliation steps.

Outputs

Parameter	Type	Description
qa_flags	GeoPackage	QA flag layer identifying records that failed yield telemetry checks.
telemetry_qc_points	optional GeoPackage	Optional point layer of telemetry QC diagnostics by record.
clean_points	GeoPackage	Cleaned yield points after telemetry and statistical filtering.
clean_map	GeoPackage	Cartography-ready cleaned yield layer for rapid map production.
confidence_points	GeoPackage	Point-level confidence diagnostics for cleaned yield records.
pass_lines	GeoPackage	Harvest pass line features inferred from telemetry trajectories.
pass_points	GeoPackage	Harvest pass points used in overlap and reconciliation analyses.
lag_corrected_points	optional GeoPackage	Optional points after lag-correction adjustment.
moisture_adjusted_points	optional GeoPackage	Optional points after moisture normalization adjustments.
filtered_points	GeoPackage	Points retained after standard filtering criteria are applied.
robust_filtered_points	optional GeoPackage	Optional points retained by robust filtering mode diagnostics.
normalized_points	GeoPackage	Yield points after normalization to common analytical basis.
reconciled_points	optional GeoPackage	Optional reconciled points after overlap and pass harmonization.
summary	JSON	Machine-readable summary report containing run metadata, QA diagnostics, and key metrics.
html_report	HTML	Human-readable customer-facing report generated from the summary contract for stakeholder review and QA traceability.

Python Example

import whitebox_workflows as wbw

wbe = wbw.WbEnvironment(include_pro=True, tier="pro")

result = wbe.yield_data_conditioning_and_qa(
    input="data/yield_points.gpkg",
    yield_field_name="YIELD",
    header_field_name="HEADER",
    profile="balanced",
    swath_width=6.096,
    output_prefix="output/yield_pipeline",
)

print(result)

License Notice

Use of this function requires a license for Whitebox Workflows Professional (WbW-Pro). Please visit www.whiteboxgeo.com to purchase a license.

Project Links

WbW Homepage User Manual Learn More