Construction Permitting Critical Path Delay Predictor
Complex adaptive reuse and conversion projects encounter long permitting bottlenecks across multiple municipal departments. Build a network graph optimization pipeline using NetworkX and LightGBM to identify critical path delay bottlenecks in municipal permitting schedules. Achieve 85%+ accuracy in predicting total permit review lead times.
What you are building
Develop a schedule network graph and delay prediction pipeline using NetworkX and LightGBM to reduce permitting friction.
How work is evaluated
Evaluates critical path calculation precision and total duration prediction error.
Datasets
Public datasets, plus the private datasets you have access to.
How submissions are scored
Critical Path Alignment
Critical path calculation must equal sum of longest dependency chain
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Lead Time Prediction MAE
Mean Absolute Error in days compared to real municipal review outcomes • target: 2.5 • range: 0-10
This dimension contributes its full weight only when the submission satisfies the requirement. Partial credit is not awarded.
Learning goals
Model municipal department permit review dependencies as Directed Acyclic Graphs (DAGs)
Train LightGBM regression models to predict department-level review durations
Perform Critical Path Method (CPM) graph traversals using NetworkX to identify bottlenecks
Output schedule risk reports highlighting high-delay dependency nodes
Resources
Historical construction permitting logs showing agency review durations and dependency sequences.
Challenge setup
Evaluates critical path calculation precision and total duration prediction error.
Challenge input
JSON with department_nodes, dependency_edges, estimated_days
NetworkX
Executes DAG topological sorting and critical path computations
Evaluated output
JSON with total_lead_time_days, critical_path_nodes
- Critical path calculation must equal sum of longest dependency chain
- Mean Absolute Error in days compared to real municipal review outcomes • target: 2.5 • range: 0-10
- Benchmark: GraphCPMBench
- Lead Time Prediction MAE target: 2.5
- 1 public reference case
- Python execution harness
- Python sandbox (unavailable on Versalist)
Shows the first task, up to three tools, and up to three evidence items.
View technical recipe
Configured tools
- Exa · Required
- Ellipsis · Optional
- AutoGluon · Optional
Evaluation contract
- Critical Path Alignment · Weight 1
- Lead Time Prediction MAE · Weight 1
Metrics
- Lead Time Prediction MAE · Target 2.5
Benchmarks
- GraphCPMBench
- general
Reference cases
- 1 public reference case
Workflow tasks
- Task 1
Input: JSON with department_nodes, dependency_edges, estimated_days
Output: JSON with total_lead_time_days, critical_path_nodes
Execution harness: Python
Evaluation tools
- NetworkX · Graph Engine
Recipe freeze marker
No recipe freeze time is recorded.