Imagine trying to build a scale model of an 11-story building. Now imagine doing it without a single floor plan, architectural drawing, cross-section, or published dimension. No CAD files. No construction documents. No helpful article in an architecture journal with the building's footprint printed to scale. Nothing.
That's the situation with the IMS Pagoda. Despite being one of the most photographed structures in motorsport — the centerpiece of the Indianapolis Motor Speedway, visible in every television broadcast of the Indy 500 — the Pagoda's architectural plans are not publicly available. I've searched. Extensively. If they exist in a public archive, I haven't found them, and neither has anyone I've asked.
So the ground floor footprint — the most critical dimension in the entire build, the foundation that every upper floor's setback is measured relative to — had to be calculated from photographs. This is the story of how.
The technique is straightforward in concept, painstaking in execution. Here's the principle: if you can identify an object of known real-world size in a photograph that also shows the building, you can calculate the building's dimensions from the photograph's pixel measurements.
The process works like this. Find a photograph where a pace car is parked near the base of the Pagoda. Measure the car's length in pixels. You know the real car is approximately 15–16 feet long. Divide the pixel count by the known length to get a pixels-per-foot ratio for that photograph. Now measure the building's width in pixels in the same photograph. Divide by the pixels-per-foot ratio. The result is the building's width in feet. Convert feet to studs at 1:38 scale — which is one-to-one — and you have the building's width in studs.
Simple math. But the execution requires care. The car and the building must be at approximately the same distance from the camera, or perspective distortion will skew the calculation. The photograph must be taken at a relatively straight-on angle, not from an extreme perspective. And the car must be fully visible — no partial occlusion that would make the pixel measurement inaccurate.
This is why 150+ reference photos matter. Most photographs don't meet all those criteria. You're looking for the handful that do — the shots where a known-size object sits near the building at a similar camera distance, in a relatively flat perspective, with clear visibility of both the reference object and the building edge. Finding those shots is the real work. The math itself is the easy part.
The ground floor footprint isn't just the base of the building. It's the reference dimension for the entire structure. Every upper floor of the IMS Pagoda is a setback from the floor below it — each level steps inward, creating the tapered silhouette that makes the Pagoda instantly recognizable. Those setbacks are proportional relationships. The third floor is a certain number of studs narrower than the second floor. The fifth floor is a certain number of studs narrower than the fourth.
If the ground floor footprint is wrong, every setback measured relative to it is also wrong. The errors compound as you go up. A ground floor that's two studs too narrow means the upper floors are all two studs too narrow relative to where they should be — and by the time you reach the eleventh story, the cumulative effect of working from incorrect base dimensions could make the entire building's proportions unrecognizable.
Get the ground floor right, and every floor above it has a fighting chance. Get it wrong, and you're building on a foundation of compounding errors.
That's why I spent more time on the ground floor dimensions than on any other single aspect of the design. Multiple photographs. Multiple reference objects. Multiple independent calculations, each cross-checked against the others. The final dimensions were not derived from a single measurement — they were the consensus of many measurements, averaged and validated until I was confident in the numbers.
This isometric render from Stud.io shows the complete ground floor footprint. Several things are immediately visible that reflect the design decisions discussed throughout this series.
The glass facades wrap the exterior of the ground floor, rendered here in transparent elements that show through to the interior. The real Pagoda's ground level is heavily glazed — large panels of glass that give the building its modern, transparent character at street level. Capturing this in LEGO means extensive use of transparent panel elements and careful attention to the framing between glass sections.
The interior room layout is visible through those glass walls. This isn't a hollow shell. The ground floor contains distinct interior spaces — rooms, corridors, and utility areas that correspond to the functional layout of the real building. You can see the walls dividing interior spaces and the structural columns that support the upper floors.
The elevator tower rises from the center-rear of the ground floor. This is the structural spine of the building — the element that runs through all eleven stories and provides the primary vertical alignment for the modular floor separation system discussed in Part 2.
And then there's the "45" visible on the track-facing facade. In the real Pagoda, the ground floor facing the start/finish line serves a ceremonial function during race events. The "45" in this render is a reference element — a marking that corresponds to the Borg-Warner Trophy's traditional display position, representing the historical connection between the Pagoda and the race it oversees. It's the kind of detail that only matters if you've spent 26 years attending the Indy 500 and know what belongs where.
This top-down view reveals what you can't see from the outside: the interior complexity of the ground floor. Every room is defined. Every elevator shaft is placed. Every structural wall is positioned according to the best interpretation of the real building's layout derived from photograph analysis.
The pink floor plates are immediately noticeable. This is a deliberate design choice, not an aesthetic one. In Stud.io, using a distinctive color for the base floor plates serves a practical purpose: it makes the structural base layer instantly identifiable when viewing the model from any angle. During the design process, when you're rotating the model and checking alignments, the pink plates tell you at a glance where the foundation is. When it comes time to order parts for the physical build, the floor plates will likely be ordered in a standard color — but during digital design, the pink serves as a visual tool.
The elevator shafts appear as rectangular voids in the floor plan. These are critical structural elements. They run vertically through the entire building and serve dual purposes: they represent the real Pagoda's elevator infrastructure, and they provide the alignment channels for the modular floor stacking system. When you set one floor module on top of another, the elevator shaft walls help guide the upper floor into the correct position.
The room divisions visible in this view were some of the most challenging elements to design. Without floor plans, the interior layout was estimated from photographs shot through the Pagoda's glass facades, from video footage showing interior spaces during broadcasts, and from logical inference about how a building of this type would be partitioned. The room positions are educated approximations — close to reality, but acknowledged as interpretive rather than architecturally precise.
The ground floor plate — and every floor plate above it — follows a set of locked engineering rules that were established early in the design process and have not changed.
Large plates on the perimeter, small plates toward the center. This rule ensures structural integrity at the edges where the floor plate is most vulnerable to bending or separation. The perimeter is where canopy overhangs attach, where facade elements connect, and where the floor experiences the most lateral stress during handling. Placing the largest available plates at the edges creates the strongest possible base along these critical boundaries. Smaller plates toward the center can float more freely because they're supported on all sides by surrounding plate structure.
No 6×24 plates. This is a cost decision. The 6×24 plate is a large, useful LEGO element — but when you need dozens of them across eleven floors, the per-piece cost on BrickLink becomes prohibitive. The solution is straightforward: everywhere the design calls for a 6×24 plate, substitute a 6×16 plate plus a 6×8 plate. The combined length is identical. The structural difference is negligible — the joint between the two smaller plates is reinforced by the overlapping plates above. And the cost savings across the entire build is substantial.
These rules might seem overly rigid for a creative project. They're not. They're liberating. When you know the floor plate rules are locked, you don't spend design time debating plate sizes. You don't second-guess whether a 6×24 might be worth the cost in one specific location. You apply the rule, move on, and focus your creative energy on the problems that actually require creative solutions — like how to capture the Pagoda's angular glass facade in a medium that's fundamentally rectilinear.
The complete process of establishing the ground floor footprint, from first photograph to final Stud.io design, followed these steps:
Step 1: Collect reference photographs. Not all 150+ photos are useful for dimensional analysis. Many are atmospheric shots, interior detail photos, or angles too extreme for pixel measurement. The dimensional analysis used a subset of approximately 20–30 photos that met the criteria: known-size reference object present, relatively straight-on angle, building edges clearly visible.
Step 2: Identify scale anchors. In each usable photograph, I identified the known-size object — a pace car (approximately 15–16 feet long), a person (approximately 5.5–6 feet tall), a vehicle at a known location — and measured it in pixels. This gave me the pixels-per-foot ratio for that specific photograph.
Step 3: Measure building dimensions. Using the calibrated pixel ratio, I measured the building's width, depth, and floor-to-floor height in each photograph. Each photo yielded a slightly different measurement due to angle variations, camera lens distortion, and imprecision in identifying exact building edges. I recorded all measurements and looked for convergence.
Step 4: Average and validate. The final dimensions are averages of the converging measurements, rounded to the nearest stud. Where measurements from different photos disagreed, I gave more weight to photos with clearer reference objects and more direct angles. The result was a set of dimensions I'm confident in to within one or two studs of the real building's proportions.
Step 5: Build the 3D reference model. Before committing to Stud.io, I built a simplified 3D model using Three.js to capture the overall massing — the building's outline, setbacks, and canopy overhangs in three dimensions. This model let me verify that the ground floor dimensions produced the correct visual proportions when extended up through all eleven stories. If the ground floor was wrong, the 3D massing model would have revealed it immediately.
Step 6: Design in Stud.io. Only after the dimensions were validated against both the photograph analysis and the 3D massing model did I open Stud.io and begin placing bricks. The ground floor was designed brick by brick, plate by plate, following the locked floor plate rules. Every dimension traces back to the pixel calculations from step 3, converted through the 1:38 scale ratio established in Part 3.
The process sounds methodical because it is. There's no room for guessing when the ground floor determines the proportions of everything above it. Every stud count was calculated, not estimated.
The ground floor footprint, as rendered in Stud.io and shown in the images above, represents months of analysis compressed into a single floor plate. It's the most thoroughly validated single layer in the entire build. Every dimension was derived from photographs, checked against multiple reference sources, and verified in a 3D massing model before being committed to the digital design.
With the ground floor locked, every subsequent floor has a defined starting point. The second floor's setback is measured from the ground floor's edges. The third floor's setback is measured from the second floor's. All the way up. The proportional relationships cascade upward through the building, each one anchored to the ground floor dimensions established through this pixel-measurement process.
Is it perfect? No. Without actual architectural plans, there's inherent uncertainty in every measurement. The ground floor could be off by a stud or two in any direction. But the methodology is sound, the cross-checks are consistent, and the visual result — when you look at the model from the same angles as reference photographs of the real building — is convincing. The proportions read correctly. The building looks like the Pagoda.
That's the goal. Not mathematical perfection down to the millimeter, but architectural accuracy that captures the building's character. The Pagoda is a distinctive building with a distinctive silhouette. If the model's silhouette matches the real building's silhouette when viewed from corresponding angles, the ground floor dimensions are close enough.
In Part 5: The Details, I'll move from the macro to the micro — the individual design challenges that make this build demanding. Canopy overhangs that need to look like they're floating. Angled glass facades that resist LEGO's rectilinear geometry. Bleacher seating that must be structurally sound at 4 rows, 19 studs deep, with 3/5/7/9 brick heights. And the open questions that still don't have answers.