Technical methods behind the architectural work.

The computational process began in QGIS and PyQGIS, where site geometry, reference points, axial relationships, and comparative reconstruction overlays were consolidated into WGS 84 / UTM Zone 36N, EPSG:32636. This projected coordinate system allowed the study to work in meters and maintain consistency across distance calculations, line intersections, centroid testing, Monte Carlo perturbation, and sensitivity analysis.

The core spatial model was organized around an east-west axial framework anchored to the Eastern / Golden Gate condition. From this datum, the research tested relationships between the Temple Mount enclosure, the Golden Gate, the Dome of the Rock, the Dome of the Spirits, and derived north-south and east-west control lines. These relationships were treated as measurable constraints capable of being formalized and challenged computationally.

The workflow moved through data preparation and layer normalization, observed-model construction, Monte Carlo uncertainty modeling, null-orientation comparison, and leave-one-out sensitivity testing. Spatial layers were cleaned and consolidated into a controlled GeoPackage, observed axial and centroid relationships were built from measured geometry, and then controlled perturbation, randomized orientation tests, and constraint-removal studies tested whether the convergence zone remained stable under uncertainty.

A key methodological decision was the use of dual-environment validation. The primary spatial construction and exploratory testing were developed in QGIS/PyQGIS, but the prepared dataset was copied into a separate Python repository and retested independently. This reduced the risk that results were artifacts of an active GIS session, manual layer handling, or software-specific project conditions.

The computational output identified a tightly constrained candidate region of architectural consistency. The result was not interpreted as historical certainty, but as a statistically bounded spatial finding: a zone where multiple independent architectural constraints converged more tightly than randomized orientation models.

The final stage translated the computational results back into architectural design. Parametric cubit-based overlays, temple footprint studies, altar offsets, processional axes, and sacred hierarchy diagrams were generated from the tested spatial anchor, allowing the project to move from analysis into design without abandoning the evidence that produced the candidate zone.

Using GIS-based analysis, Monte Carlo simulation, geometric constraint testing, and computational archaeology workflows, the project seeks to identify recurring organizational systems embedded within temple precincts, civic compounds, and ceremonial landscapes throughout the Ancient Near East.

Rather than approaching historical architecture solely through textual interpretation or stylistic reconstruction, the research treats architecture as a measurable spatial system shaped by procession, hierarchy, visibility, orientation, environmental logic, and ritual order.

Particular attention is given to how axial geometries persist through periods of reconstruction, expansion, political transition, and stratigraphic change, allowing later architectural phases to be studied as transformations of earlier spatial orders rather than isolated formal episodes.

The framework combines probabilistic spatial analysis, topological persistence studies, centroid and alignment testing, computational reconstruction methods, and architectural design translation.

Current investigations include comparative analysis between Levantine, Israelite, Mesopotamian, and broader regional precinct systems to evaluate whether measurable spatial signatures emerge across differing cultural and temporal contexts.

Research framework and comparative methodology development are currently in progress. Initial computational models and probabilistic spatial studies were completed through prior thesis research, with ongoing expansion into cross-site comparative analysis and generalized axial-system detection methodologies.

The system operates through a staged reasoning pipeline: user input is first parsed into extracted concepts, then evaluated against a custom architectural kernel containing weighted heuristics, precedent anchors, vocabularies, and prompt rules. These activated priors guide the model toward grounded architectural reasoning, producing structured context, design constraints, implicit signals, and a narrative brief that can inform schematic design decisions.

A second test case examined a compact 1,800-square-foot urban infill residence in Austin, Texas, located beside a commuter rail line and public alley. The engine extracted conditions such as zero-lot-line constraints, rail adjacency, acoustic isolation, rooftop outdoor space, passive solar shading, and lock-and-leave security. It then activated relevant priors for solar exposure, acoustic buffering, narrow-lot sectional thinking, and precedent-based mass/slot logic.

The generated design narrative translated these constraints into architectural moves: a rail-facing acoustic wall, a service spine functioning as a noise gradient, a sectional solar strategy, a compressed rooftop outdoor room, and an alley wall conceived as a security threshold. The result demonstrates the system's core purpose: not to produce a final design, but to reveal the latent architectural logic embedded within a client/site brief.

At its current stage, Context_Engine functions as a proof-of-concept for semantic constraint modeling in architecture. Future development will expand the system toward RAG-supported building code, zoning, accessibility, and technical specification parsing, allowing architectural briefs to be cross-referenced against compliance logic and source-grounded requirements.

Rather than treating a floor plan as a purely graphic composition, the project frames layout design as a computational problem: spaces are understood as nodes, relationships are treated as edges, and architectural constraints become rules that can be tested, sorted, and visualized.

At its core, the tool converts architectural programming data into a structured spatial network. Each space is assigned attributes such as size, function, access priority, privacy level, adjacency needs, and circulation requirements, then organized into a logical arrangement based on those relationships.

The system includes a visualization engine that maps the physical connections between spaces, showing how rooms relate through direct adjacency, circulation sequence, public/private hierarchy, and programmatic dependency. These diagrams reveal whether a layout is functioning coherently before it is developed into a full architectural plan.

Highly connected public spaces can be positioned as primary anchors, while service areas, private rooms, and support spaces are placed according to dependency and access logic. The resulting output becomes a computationally informed planning diagram that can guide early schematic design.

The tool supports custom topological sorting, spatial adjacency mapping, physical and functional connection visualization, public/private hierarchy, circulation logic, early-stage layout optimization, and diagrammatic comparison between multiple planning scenarios.

By translating program requirements into computational logic, the tool creates a bridge between architectural intuition and algorithmic analysis. It helps designers ask which spaces must connect, which relationships are optional, which adjacencies create conflict, and which arrangement produces the clearest spatial order.