Managing Jumper Connectors GIS Cleanup with Python and AI
Managing Jumper Connectors GIS Cleanup with Python and AI
Abstract
The City of Naperville Electric Utility is migrating from ArcMap / ArcFM to the Esri Utility Network. A pre-migration requirement was cleaning up features before they become problems in the new data model. One of the most significant was around 9,000 Jumper connectors connecting underground cables to transformer banks — a fix that would have taken approximately two full-time employees 12 weeks manually. Instead, the GIS team completed it in a fraction of that time through structured analysis, Python automation, and AI-assisted development using Claude (Anthropic). The presentation tells the full project story. It began with proof-of-concept testing on a File Geodatabase copy of the production network, where a script ran cleanly in about one hour. Moving to a versioned SDE environment with more than 250 active project versions introduced a completely different class of problems: workspace transaction mode conflicts, an incompatibility between the two ArcPy editing APIs that prevented deletion and snapping in the same edit session, zero-length geometry errors from incorrect operation ordering, and a performance collapse from 1 hour to 240 hours across 15 substation runs. Resolving these required a two-phase edit session redesign and AI-assisted performance tuning that brought total runtime down to 29 hours. The result: fewer than 100 version conflicts in three reconcile passes and a clean feeder report across all 16 substations. Attendees will learn practical lessons about geometric network editing in versioned SDE environments, the real role AI played as a development collaborator, and how to approach large-scale data quality projects with confidence.