Suburban Land Assemblage Underwriter with AutoGen and Cognee Knowledge Graphs
Faith-based institutional capital is executing long-horizon suburban land plays requiring complex parcel ownership graph mapping. Build an AutoGen multi-agent system combined with Cognee graph memory to automatically identify optimal land assemblage targets with 92% relational accuracy.
What you are building
The core problem, expected build, and operating context for this challenge.
Construct an AutoGen agent network backed by Cognee dynamic knowledge graph memory to evaluate contiguous suburban parcels and compute yield-adjusted land acquisition strategies.
How work is evaluated
Evaluates AutoGen and Cognee output on suburban parcel group selection and total assemblage yield calculation.
Challenge datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
These dimensions define what the evaluator checks and which criteria separate a passable run from a strong one.
Cluster Continuity Check
Verifies that returned parcel cluster contains valid contiguous parcels.
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
acreage_calculation_accuracy
Percentage accuracy of summed total acreage across assembled parcels • target: 1 • range: 0-1
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
What you should walk away with
Implement AutoGen agents (ParcelAnalyst, ZoningAdvisor, FinancialUnderwriter) collaborating in a group chat.
Leverage Cognee to vectorize land registry records and build graph relations between adjacent parcels.
Calculate aggregated acreage, assembly premium cost models, and target IRR.
Generate automated acquisition proposals for long-horizon capital deployment.
Reference links and supporting material
Dataset of 60 suburban land parcel networks with topological adjacency matrices, current tax assessed values, zoning designations, and acquisition cost histories.
Challenge setup
Evaluates AutoGen and Cognee output on suburban parcel group selection and total assemblage yield calculation.
Challenge input
JSON list of parcel objects with acreage, zoning, asking_price, and adjacency_list
AutoGen
Mandatory framework for multi-agent group conversation orchestration.
Cognee
Required memory tool for extracting and querying topological land graphs.
Evaluated output
JSON with contiguous_cluster_ids, total_acreage, total_estimated_cost, and calculated_irr
- Verifies that returned parcel cluster contains valid contiguous parcels.
- Percentage accuracy of summed total acreage across assembled parcels • target: 1 • range: 0-1
- Benchmark: Multi-Agent Graph Reasoning Benchmark
- Acreage Calculation Accuracy target: 1
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Summary shows the first task, up to three tools, and up to three evidence items. Open the technical recipe to view all configured items.
View technical recipe
Configured tools
- Cognee · Required
- AutoGen · Optional
- Traceloop · Optional
Evaluation contract
- Cluster Continuity Check · Weight 1
- acreage_calculation_accuracy · Weight 1
Metrics
- acreage_calculation_accuracy · Target 1
Benchmarks
- Multi-Agent Graph Reasoning Benchmark
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON list of parcel objects with acreage, zoning, asking_price, and adjacency_list
Output: JSON with contiguous_cluster_ids, total_acreage, total_estimated_cost, and calculated_irr
Execution harness: Python
Evaluation tools
- AutoGen · Multi-Agent Framework
- Cognee · Agent System Memory
Recipe freeze marker
No recipe freeze time is recorded.