Irradiance, Shading & Production Data Have to Agree
A solar proposal is, at its core, a promise. It promises how much electricity a system will produce. Google Solar API and PVWatts integration make that promise checkable, both before and after installation.
That promise depends on three data layers. Irradiance data sets the baseline. A shading model adjusts it for the specific roof. Post-install monitoring either confirms the estimate or contradicts it.
Enphase and SolarEdge monitoring APIs verify production after installation. A permit-packet generation layer turns the same design data into a submittable package.
Verify each API’s current terms, pricing, and rate limits before making architecture decisions. This stack spans a Google Cloud product, a federal research-lab tool, and two competing monitoring ecosystems. All three evolve independently.
These integrations also have to support two very different user environments. Custom mobile app development gives field teams access to site, design, and production data while they are on location, treating irradiance accuracy, shading model output, and post-install monitoring data as connected inputs to the same proposal rather than three separate data pulls. The back-office workflows that create proposals, review system designs, manage project data, and prepare permit documentation are built for scale
Google Solar API for Roof Geometry & Irradiance
What the API Provides
Google’s Solar API is part of Google Maps Platform. It provides building-level roof geometry, usable roof area, and irradiance data. That data comes from aerial imagery and digital surface models.
This is the foundational data layer most solar design platforms build on. Aurora Solar and OpenSolar both layer their own photogrammetry and shading-analysis features on top of it. A custom platform follows the same pattern.
The API returns data at the roof-segment level, not just a single building footprint. Each segment carries its own pitch, orientation, and sun-exposure estimate across the year. The granularity lets a design tool place panels segment by segment, not on one average roof angle.
Where It Needs Supplementing
Coverage and resolution can vary by region. Complex roof geometries or dense tree cover sometimes need a supplementary review pass. Relying on raw API output alone isn’t always enough.
A manual or AI-assisted check catches what the base layer misses. Verify current API coverage, rate limits, and pricing directly before scoping a specific integration. Pricing terms shift more often than the underlying geometry data does.
NREL PVWatts for Production Estimation
NREL’s PVWatts calculator remains the standard, freely available production-estimation tool. It estimates annual energy production from system size, location, and orientation. Installers, financiers, and utilities all treat it as a trusted baseline.
A platform should use PVWatts, or an equivalent transparently documented model, as that baseline. Site-specific shading and obstruction data from the roof-modeling pipeline layers on top of it. A generic regional estimate presented as site-specific is a real accuracy problem, not a shortcut.
PVWatts inputs include system size, module type, array tilt, azimuth, and system losses like wiring and inverter inefficiency. Feeding it accurate, roof-specific values instead of default assumptions is what separates a defensible estimate from a rough guess. Small errors in tilt or azimuth compound over a full year of production. Custom software development for the production-estimation backend handles the PVWatts API integration, roof-specific input normalization pipeline, shading-adjustment layer, Enphase and SolarEdge monitoring data ingestion, and permit-packet generation workflow that tie irradiance data, design output, and post-install production into one verifiable record.
Design-Engine Photogrammetry & Shading Analysis
The photogrammetry and shading-analysis layer turns raw roof geometry and irradiance data into a usable design. Aurora Solar and OpenSolar compete in this exact category. Aurora’s own AI-powered modeling feature is branded Aurora AI.
Automated obstruction detection increasingly replaces fully manual roof tracing in this layer. A custom platform has two real paths here. It can integrate with a third-party design engine’s API where one exists. AI product development for a first-party roof-geometry and obstruction detection model handles the computer-vision model training pipeline, aerial imagery preprocessing layer, and model evaluation framework that keep shading estimates accurate across the regional satellite imagery variation found in US residential solar markets. Or it can build a first-party computer-vision model for roof and obstruction detection. That is a genuine, buildable feature, not just a description of what competitors already offer. Either path needs to hold up against the same accuracy bar.
Shading-analysis accuracy across a full year is what actually matters, not a single snapshot. That accuracy determines whether the eventual production-monitoring data confirms the original proposal or contradicts it. A mismatch there erodes homeowner trust after the sale. How roof-modeling automation, financing comparison, permit-packet generation, monitoring integration, CRM workflow, and field-sales features connect into the complete solar proposal platform feature architecture runs through Solar Proposal Software Features: Must-Haves for a US Residential Solar Sales & System Design Platform in 2026.
A reasonable QA step compares the shading model’s output against a handful of completed installs before it ships to production. A model that overstates production on shaded roofs needs fixing before the next sales conversation, not after a support ticket.
Enphase & SolarEdge Monitoring API Integration
Enphase and SolarEdge are the two dominant residential solar inverters and monitoring ecosystems in the US. Each exposes its own monitoring API. Enphase reports data per panel, while SolarEdge reports per system, in near-real time.
The platform should integrate with whichever ecosystem a given installation actually uses. Actual production data feeds back into the original design record. That closes the loop between what the proposal promised and what the system delivers.
This data supports performance-guarantee comparison, warranty support, and underperformance alerting. For third-party-owned lease or PPA systems specifically, this data often feeds the financing partner’s own reporting and billing systems.
Alert thresholds need to account for normal seasonal variation, not just a flat daily target. A system producing less in December than July isn’t underperforming; it’s following the sun. Alerts tuned only to a flat baseline generate false positives that eventually get ignored.
Since TPO financing makes up a meaningful share of many installers’ businesses, this seasonal-alert logic is worth scoping as its own integration requirement, not an afterthought. iOS app development for the monitoring dashboard configures APNs push notifications for underperformance alerts, delivers production data from Enphase and SolarEdge APIs through an offline-capable data layer for areas with limited connectivity, and submits an App Store privacy nutrition label disclosing customer address and production data collection before the submission goes into review.
Roof-modeling and monitoring integration complexity are primary cost drivers for this kind of build. How roof-modeling automation complexity, post-ITC financing comparison logic, AHJ-specific permit packet automation, Enphase and SolarEdge monitoring integration, and multi-office CRM scope each affect the investment range across proposal MVP, full sales and design platform, and enterprise EPC platform tiers runs through Cost to Build a Custom Solar Sales, Design & Proposal Platform for a US Solar Installer: Full Budget Breakdown.
Permit-Packet Generation & AHJ Submission
The same design data that generates the proposal should also generate the permit packet. One-line and three-line electrical diagrams come from that same source. Structural attachment details and NEC-690-compliant labeling follow the same path.
Formatting needs to match the specific AHJ’s submission requirements, not a generic national template. Code-edition adoption and submission formats both vary by jurisdiction. This layer should be built as AHJ-configurable rather than hard-coded. The solar design platform and permit dashboard where installers manage AHJ-specific packet formatting, track submission status across jurisdictions, configure NEC-edition templates, and review one-line and three-line diagram outputs require web application development built around role-based access, configurable permit-packet export formats, and real-time permitting queue visibility.
A permit-packet generator tuned to one city’s requirements will get rejected in the next one over. Configurability here isn’t a nice-to-have. It keeps the platform usable as the installer expands into new markets.
Version control also matters here. A design can change after initial permit submission, whether from a shading correction or a homeowner request. The packet needs to regenerate cleanly, without a manual redraft of every diagram and label. Android app development for the in-field solar sales app handles offline permit-packet data capture for areas with limited connectivity, FCM push notifications for AHJ submission status updates and design-change alerts, and Google Play data safety disclosures covering customer address and financial information before the submission goes into review
Final Thoughts
Founders who treat irradiance data, shading analysis, and post-install monitoring as one continuous pipeline ship something durable. Three disconnected tools produce a proposal that looks good on paper and falls apart after install. A self-verifying pipeline doesn’t.
If irradiance and shading accuracy are the core of your platform, scope this integration stack deliberately. Let the shading model and the post-install data check each other. That separates a platform homeowners trust from one that generates a support ticket six months later. To see how an AI software development company approaches Google Solar API and PVWatts irradiance pipeline design, automated roof-geometry and obstruction detection model development, Enphase and SolarEdge monitoring API integration, AHJ-configurable permit packet generation, and seasonal-alert logic for TPO financing portfolios for US residential solar installers and EPC contractors, explore our work with solar technology development teams
