This model turns a limited historical workbook exercise into a visual scenario tool with maps, tables, filters, and PDF-ready summaries.

The earlier Excel workflow is useful as a learning and historical reference, not as a current data authority. We do not know whether its geography or portfolio data is complete or current.

This version keeps the tabular review discipline while adding map symbology, hub coverage circles, no-tech screening, scenario controls, and print-ready graphics.

Gold circles and yellow spokes show the recommended hub coverage footprint. Blue circles emphasize the selected hub and its assigned properties.

This is more complex than picking the nearest office. The model is trying to balance coverage, staffing, travel, and facility responsibility at the same time.

Drive-time threshold: wider coverage can reduce hub count, but may add paid travel time and reduce productive work hours.

Current satellite symbology: property satellite size uses three rentable-SF bands, and satellite color uses owner type: owned, leased, licensed, or other.

Needs-tech screening: properties under 1,000 rentable SF default to no tech needed. The master property table checkbox can override that default for testing.

Property mix: offices, warehouses, parking lots, land, antennas, hospitals, and military-base space do not create the same maintenance workload.

Ownership and responsibility: owned, leased, licensed, shared direct, and chartered sites may require different levels of ARC facilities support.

Workload capacity: max properties per person is a placeholder for future workload hours, not a claim that all facilities take equal effort.

Division boundaries: division filters focus a scenario, but the model should not assume technicians can never support a nearby property across a division line.

Routing quality: straight-line distance is useful for screening, but road networks, traffic, water crossings, mountains, parking, and access constraints can change the answer.

Scenario cost: fewer recommended people can still produce a weaker operating model if the remaining people spend too much paid time driving.

Choose a drive-time threshold to adjust the coverage radius.

Filter by division or region to focus the map and results on that operating area.

Select a recommended hub on the map or in the table to zoom to its radius circle and assigned properties.

Use Download hub summary for spreadsheet review or Print report / PDF for a clean executive summary.

The current bundled dataset is the app's working portfolio sample. The workbook is preserved only to compare old spreadsheet assumptions against the current visual model.

This exercise uses bundled facility data so the model opens directly. A future version will let users upload the current ARC real estate portfolio for fresh analysis.

Facilities include a wide variety of property types, from offices and warehouses to parking lots, land, antennas, and space within military bases.

The current model uses a straight-line fallback, so it is useful for quick screening but should not be treated as final road routing.

Check ArcGIS first because ARC may already have Esri routing privileges or credits. If credit costs are unfavorable, use a self-hosted OSRM or Valhalla server for predictable monthly infrastructure cost.

Save every road-routing result in a permanent route cache. Threshold changes should reuse cached drive times and rerun the solver, not call the provider again.

Use focused calibration samples to compare straight-line estimates against road drive times. Southern California is a good stress test because congestion, freeway access, mountains, and coastal bottlenecks can make same-mileage trips behave very differently.

Until true routing is connected, a regional travel efficiency factor can make the straight-line fallback more honest by shrinking or expanding effective coverage by geography.

Dense congestion areas such as Los Angeles may use factors like 0.60–0.75, while open-road areas such as West Texas may use 1.00–1.10.

Regions with water or mountain barriers, ferries, sparse crossings, or indirect roads should receive lower factors even when they are not dense metros.

A later version should allow compensation by region so each recommended person can be costed with the correct local pay assumption.

Future regional pay rates should be editable scenario inputs, separate from hub selection and drive-time coverage logic.

Use fully loaded cost, not salary alone. A loaded labor factor accounts for benefits, payroll taxes, insurance, retirement, HR/admin burden, and field support costs.

Because this model involves driving, convert annual loaded cost using productive service hours available, not just salary divided by 2,080 hours.

Factors to consider include driving time, admin time, parts runs, vendor coordination, meetings, leave, training, and equipment or vehicle costs.

For remote or low-hour sites, compare internal loaded labor plus travel against vendor cost so the model can flag outsource candidates instead of forcing inefficient internal coverage.

The report should show staffing cost by region and total modeled staffing cost, while still keeping recommended hubs separate from recommended people.

A later version should grade each facility by estimated service-tech workload hours per week or month instead of treating every property as equal.

Maintain a service-tech decision ledger keyed by building ID so quarterly or monthly portfolio uploads can reapply reviewed decisions and flag new or changed properties.

Future satellite size and color should come from a need score that blends rentable SF, site type, building type, age, usage, alternative activity signals, ownership responsibility, B&O intensity, and strategic importance.

Ownership Factor should make licensed/free-space locations low, leased sites moderate, owned sites high, and owned-building/leased-land sites highest because ARC carries more facility responsibility.

Site Type Factor should raise Biomed Direct, Shared Direct, and NHQ Direct sites above Indirect sites because operating intensity and facility expectations are higher.

Building Type Factor should keep parking lots, land, antennas, military-base space, and simple storage low while raising office, warehouse, industrial, hospital, and headquarters-style facilities.

Market Tier Factor and Strategy Factor should capture local complexity and real estate intent; Tier 1 markets and Analysis sites can be slightly higher, while Exit or N/A sites can be lower unless manually overridden.

B&O Intensity Factor should use FY26 B&O / SF where available, with higher operating cost per square foot increasing estimated workload.

Recommended people should eventually be calculated this way: divide total assigned weekly hours by max service hours per person per week, then round up to the next whole person; default: 32 productive service hours per person per week.

For outsource candidates, compare internal loaded cost plus travel against vendor cost, including monthly minimums, hourly rates, trip charges, and remote premiums.

Brivo security check-ins and network usage can help estimate building activity, but they should be supporting signals, not standalone staffing rules.

Brivo check-ins can show staff or visitor activity patterns, training days, after-hours use, and whether a site is regularly occupied.

Network usage can show active digital use. Normalize sent and received totals into GB per day and GB per rentable SF before comparing sites.

Example: high Brivo check-ins plus high network usage likely points to an active staffed site that deserves higher review weight.

Example: high network usage with low Brivo check-ins may indicate servers, backups, phones, cameras, or other infrastructure rather than occupied workspace.

Example: high Brivo check-ins with low network usage may indicate field staff touchpoints, short visits, training setup, or a site where network data is incomplete.

Example: low Brivo and low network usage at a warehouse, parking lot, land parcel, military-base room, or disaster cache should not automatically mean no service-tech support.

Use simple tiers such as Low, Moderate, High, and Infrastructure-heavy until these signals are validated against occupancy, work orders, and known site functions.

Alternative Activity Factor should start as a small multiplier or review flag, with manual override and an override reason kept in the decision ledger.

GeoJSON gives GIS teams a map-ready file they can load into ArcGIS, QGIS, Mapbox, or other spatial tools with facility coordinates and model coverage attributes attached.

Batch only missing route pairs, show cache progress in the UI, and reuse cached drive times when only 2–3% of facilities change.

Max properties per person converts overloaded hub assignments into multiple recommended people at the same hub; it does not yet run a full workload-balanced optimization.

ARC Real Estate / Facilities Strategy contact