Example Of Site Analysis In Architecture

Site analysis isn't a box you check before design actually starts

I've been doing this for long enough that I can usually tell whether a firm's site analysis phase was real or just something their admin assistant compiled from Zillow and Google Street View in an afternoon. The difference shows up later, during construction administration, when someone realizes the south facade they drew isn't the one the sun actually hits at 2pm in November. The method is straightforward enough that beginners tend to breeze through it, which is the problem. You go to the site, you measure things, you note what's there, and then you translate that into design decisions. Most people stop there. The people who don't stop there are the ones who still get called back to fix drainage problems two years after the building was occupied.

Example Of Site Analysis In Architecture

Here's a concrete walkthrough. A few years back I was consulting on a residential renovation in a coastal Massachusetts town. The project was a simple kitchen expansion, nothing heroic. The original drawings showed the back of the house as a blank canvas with a 12-foot by 16-foot addition going up. What the site analysis missed was the seasonal stormwater pooling pattern in the yard behind the proposed footprint. I walked the site at 7am after three days of rain and found standing water roughly where the addition's foundation line fell. The surveyor had come during a dry week and marked the property lines correctly but didn't note the hydrology. We shifted the addition six feet west, changed the footing detail to a reinforced grade beam, and avoided having the future homeowner explain to her inspector why his new kitchen floor was two inches lower than the rest of the house. That's the kind of thing site analysis catches. It's not glamorous. It also saves people money.

What site analysis actually covers

The components break down into about six categories, though nobody does all of them perfectly: Topography — slope direction, contour intervals, cut and fill areas, existing grade relationships to adjacent properties. You need this for drainage, foundation type, and access. Contour maps from the survey are a starting point, not the finish line. A survey typically has 1-foot contours; you should be walking the site and noting micro-drainage paths that won't show up on that map. Solar and wind — sun path at the specific latitude, prevailing wind direction, shading from adjacent structures and mature vegetation. Most firms use SunTools or a dedicated app for this. The apps are fine for initial orientation but they won't account for a proposed tree that gets planted five years from now. I always overlay the solar diagram on a site photograph taken from each cardinal direction so you can see what's actually blocking light, not what the software assumes is there.

Vehicular and pedestrian circulation — existing driveway positions, sidewalk connections, neighborhood pedestrian flow, delivery and service access points. This determines where the main entrance goes, where the garage fits, and whether you've accidentally designed a building that requires residents to walk through someone else's yard to get to their car. Context and adjacent buildings — height, massing, setback, material palette, architectural style of neighboring structures. This is the part that turns into style policing most often. The useful reading here is about scale and rhythm, not about copying clapboard siding because the house next door has it. Utilities and infrastructure — location of water, sewer, electrical, gas, telecom connections. Some municipalities have maps. Most don't have accurate maps. Field verification by calling 811 and watching the mark-out crew is the only way to know for sure where the existing lines are. I've designed three buildings on paper over utility easements that weren't shown on any map until the excavator hit them.

Vegetation and ecological features — mature trees to preserve, wetlands, habitat areas, erosion-sensitive zones. Local zoning often has tree protection ordinances that carry real penalties. One client in Connecticut got fined forty-two thousand dollars for removing an oak that was marked on the arborist report but not flagged in the field. The arborist had staked it. No one looked at the staked tree.

How to actually do it without wasting three days

Start with the paperwork. Pull the survey, the zoning map, the tax assessor's parcel data, and any environmental overlays from the municipality's GIS portal. This takes maybe twenty minutes online if the town's database isn't completely broken. Then go to the site. Bring a laser measure, a compass, a camera, and a notebook. Not all at once, obviously, but those are the tools. Walk the perimeter first. Note the street frontage, the setbacks as they currently exist, any irregularities in the lot line. Then move inward. Photograph every cardinal direction from the center of the lot. Note the slope by standing at different points and looking across. You'll feel it in your legs before you measure it with a device. For solar orientation, take your photos at roughly the same time of day over two different periods if possible — late spring and early fall give you a good range. Mark where shadows fall at those times. Your shadow on the ground tells you more about winter overshadowing than any diagram does.

Check the neighborhood. Drive it. Park somewhere and sit for ten minutes. Watch how people move, where cars turn, where the mailboxes are positioned. The details that matter most — the sidewalk gap, the driveway apron that's been cut into the curb, the spot where neighbors leave trash cans — only show up when you're actually present.

Common mistakes that cost real money

The biggest one is assuming the survey is complete. It isn't. Surveys show property lines and improved structures. They rarely show grading changes made over time, underground drainage paths, or the actual mature canopy of trees that may have been removed and replanted since the survey was done. Another is relying entirely on sun path software without verifying on site. The software assumes a clear horizon. Most sites don't have a clear horizon. A rooftop HVAC unit two blocks away can cast a meaningful shadow in winter if the geometry aligns, and the software won't know about it unless you feed it the right obstacles. A third is ignoring the existing conditions that aren't physical. Noise sources, odors, visual distractions from nearby roads or parking lots — these affect window placement and room function. I once designed a bedroom window that opened directly toward a high school football practice field. The client loved the natural light. She did not love the 6am whistle for three years.

There's also the temptation to skip the utility check because the plans look clean on paper. This is how you learn where the septic tank is by accident.

Tools I actually use versus tools I've abandoned

For field work: a Bosch laser measure, a Brunton pocket transit, a DSLR or mirrorless camera with a wide lens, and a physical notebook. The Brunton gives me bearing and slope angle in one tool. The laser measure covers distances up to 300 meters. The camera is for documentation and for creating the photo-panels I layer with diagrams later. For office work: Rhino or SketchUp for modeling the site context, Grasshopper with Ladybug for solar analysis when the project needs precision, and ArcGIS or QGIS for zoning and environmental overlay work. QGIS is free and handles municipal GIS shapefiles without complaint. If your firm budget allows it, ArcGIS is faster but the marginal gain rarely justifies the license cost for residential-scale work. I used to swear by AutoCAD's site plan templates. They're rigid and they encourage drawing the wrong level of detail. Now I start most site analyses in SketchUp because the 3D context model makes it immediately obvious when a proposed massing violates a setback or casts an unintended shadow on a neighbor's yard.

When site analysis fails and what to do instead

Site analysis breaks down in three scenarios. The first is a site with no existing records at all — no survey, no As-Builts, no tax map that matches current boundaries. This happens more often than you'd think with older rural properties. In those cases, hire a licensed surveyor first. Everything else waits for that. You can't analyze what you can't legally locate. The second scenario is a site where the surrounding context is in active transition. A new highway is being built three lots over. A commercial zone is being rezoned. The existing conditions you document today may be irrelevant within eighteen months. In that case, the site analysis needs to include a forward-looking component — zoning change applications, transportation studies, anything publicly filed. The analysis becomes partly a research exercise rather than purely an observational one. The third scenario is when the site is so constrained or so problematic that the analysis confirms there is no workable solution without major intervention. A steep slope over 30 percent, a floodplain designation, a protected species habitat. The honest output of the analysis is sometimes a recommendation not to build where the client wants to build. That's a valid result. I've had clients initially resist it, then thank me later when they avoided a permitting nightmare that would have stalled the project for two years.

Deliverables that actually get used

A well-done site analysis package includes: a base map with existing conditions overlaid, a solar diagram showing summer and winter sun paths, a shadow study at key dates, a drainage and grading assessment, a circulation map showing existing and proposed routes, a vegetation inventory with preservation recommendations, a utilities locating note, and a set of context photographs with diagrams pinned to them. That's roughly 15 to 25 pages for a typical residential project. The photographs with diagrams are the part that matters most during design development. When you're arguing with a structural engineer about where to place a column, a photo showing the existing tree root zone at that exact location is worth more than a verbal description. When you're presenting to a client who doesn't understand why the window can't go where they want it, the shadow diagram on the photo is the easiest thing to point at.

A practical note on timing

For a standard residential project, a competent site analysis takes about 6 to 10 hours of field work plus another 10 to 15 hours of office processing and diagram production. That's two to three days of real work. Some firms try to compress this into a single afternoon, and the resulting drawings are usually adequate for permit submission but inadequate for design decisions that matter. If you're designing a building that will sit on that site for forty years, fifteen hours is a small investment. The firms that skip it often find out during the bidding phase when a contractor flags a condition that wasn't documented, or during construction when the foundation detail doesn't match the actual soil profile. Those moments are expensive. I've also seen the opposite extreme — site analysis so exhaustive that it becomes paralysis. I worked with a firm that spent six weeks on a site analysis for a small library addition. They mapped every species of plant, measured every shadow angle, modeled the wind tunnel. The final design decisions were based on three observations: the sun comes from the south, the driveway is on the east, and the view they wanted to preserve was toward the water. The other fifty-seven pages were impressive and largely unused.

There's a balance. The analysis should inform the design, not replace it. If you find yourself generating diagrams that no one references after week one, you've gone too far.