Building a Spatial Framework at The Ohio State University from Bedrock Up

TreavonClark

Managing smart campus operations at The Ohio State University is no small feat. Across a portfolio of 773 BIM (Building Information Model), 42.7 million gross square feet of space, and 14,500 acres of land, there are countless use cases to delve into.

For the past five years, I have been a GIS Application Developer and a member of the Facilities Information and Technology Services (FITS) department. In that time, FITS has positioned itself as the keystone of a massive geospatial ecosystem. In this blog post, I want to share my perspective on our spatial framework, including some of our successes and the challenges we’ve faced along the way.

The Three Legs of the Stool

FITS is comprised of three key areas of specialization:

Architectural Services: This team maintains facility floor plans through Revit. These BIM serve as our system of record for architectural information. The architectural models inform our GIS.
Data Services: This team maintains space- and building-level data in an independent record system called our Space Information and Management System (SIMS). SIMS relies on BIM and, in turn, informs our GIS.
Geospatial Services: This team maintains the campus base map and geospatial assets.

FITS operates much like the three legs of a stool, described in this blog post by John Young, Senior Consultant at Langan Digital Solutions. Historically, floor plan management and space management have worked seamlessly together. GIS is the latest to join this collaboration, and I agree with the premise that GIS is best equipped to be the "glue" that ties everything together. It is the platform that elevates visualizations, analytics, and workflows.

This three-team equilibrium gives us a holistic perspective on operations. No single team holds all the data, or all the data outflow. We have realized that this is only possible through heavy automation, an organizational structure that reinforces cross-communication, and collaboration beyond our immediate department.

While FITS does not track every asset on campus, other units manage their own specialized assets, any campus facility data can be leveraged by one of our three teams. FITS creates a spatial framework, almost like a canvas, where all campus operations and assets can be mapped and described.

Current Success: Building Upon Bedrock

Today, the cross-collaboration between our BIM, Space, and GIS data is fully automated. BIM and Space data are not simply imported and forgotten; all three act as systems of record, meaning a change in one can prompt a necessary update in the others.

Here is a practical example of how this works: GIS contains the best representation of our landscape through our base map topology, which we keep accurate by regularly reviewing drone imagery and construction plans. Let's say a routine drone flight identifies a new, permanent shed on campus.

Because this shed is not found in SIMS (our building system of record), a change is triggered. Space Management assigns the new shed a name and facility ID. Architectural Services then determines that a model is required and creates a BIM, which is shared back with both the GIS and Space systems. Timely integration and routine data audits like this are critical for keeping our information consistent across the board.

Utilizing BIM in Multiple Ways

A traditional confluence of floor plan, space, and GIS data is the ArcGIS Indoors ecosystem. For my work, the primary value of these tools lies in BIM-to-GIS conversion and asset collection integration. We use ArcGIS Field Maps and ArcGIS Survey123 for field verification, and ArcGIS Experience Builder for viewing floor plans and querying space data.

A space query showing a filter for circulation space on the first floor of all buildings in a 2D app.

However, the value of BIM within GIS truly shines when we leverage 3D scenes:

Building Scene Layer Objects: Converting a single BIM to a single SLPK and hosting it in ArcGIS allows end-users to curate their own scenes in Experience Builder. Using the "add data" widget, users can add as many BIM to a scene as their client can render.

(view in My Videos)

Global 3D Scenes: Rendering entire building scenes can be computationally heavy. To solve this, I identified the most important aspects of the BIM for our stakeholders and extracted only those layers into 3D Object layers. This provides lightning-fast access to building facades alongside a unique perspective on space data.

BIM exteriors in a single 3D scene showing a portion of the Columbus Campus.

Utilizing APIs

Once the BIM is imported to GIS, the room objects can be joined to SIMS to embed rich space information. Luckily, this is a simple JSON API to GIS feature workflow. Additionally, the space information utilizes the same unique IDs found within the room features in the BIM. This means the BIM and Space data are primed for GIS visualization, requiring no correction upon ingestion into the GIS ecosystem.

Challenges Faced

The Automation Hurdle

A prerequisite for this entire framework is reliable data. As a change is made in a system of record, it needs to be reflected in the referencing system, requiring strict automation for both BIM and Space data.

As published BIM are made available, our system runs a date comparison. If the ".rvt" file is newer than the data in our GIS, the old building is deleted, and the new one is added. All indoor asset layers remain in place, though they trigger an audit depending on the asset type. Finally, an API querying space data from SIMS embeds all matching unit information directly into the polygon’s attributes.

RVT to Indoors Script Overview

#Find .rvt filepaths
for identified_file in rvt_files_by_campus:
   #Custom function, decide if a building needs re-ran, ran for the first time ever, or skipped
   ran_before, gis_older = value_exists(...)

   #Custom function to delete the entire facility from GIS
   delete_prior_run(...)
   arcpy.conversion.ExtractBIMFileFloorplan(...)

   #Do as much cleanup as needed on the temp gdb layers generated. I am skipping a lot of cleanup.

   #Repairing geometry helped with some fringe cases.
   arcpy.management.RepairGeometry(...)

   #I highly recomend a pairwise dissolve on the linework for performance.
   #This took the DB record count from 1.2 million to ~25,000.
   arcpy.analysis.PairwiseDissolve(...)

   #When the data is prepped, append the records to your prod db
   #One append per data layer
   arcpy.management.Append(...)

   #Custom function, clean up after each building is run
   clear_gdb(scratch_DB)

RVT to 3D Shells and Rooms

#Use a similair decision process to identify if a building needs run

#If it does need run, add it to a map
m = aprx.listMaps("Map_name_here")[0]
m.addDataFromPath(rvt_to_run_path)

#When you add a BIM to a map, you get an "exterior shell" layer.
#I rely on this layer (for now) for my exterior shells.
#Find the layers of interest and append them to a gdb into multipatch layers. 
#One for all shells, one for all rooms
for layer in m.listLayers(...):
   for l in layer.listLayers():
      if r"ExteriorShell" in l.longName:
         arcpy.management.Append(...)
      if r"Room" in l.longName:
         arcpy.management.Append(...)

#Publishing this is manual for now, shells take considerable time to publish.

RVT to 3D Building Scene Layer Script Overview

#Use a similair decision process to identify if a building needs run

#Turn BIM into building scene layer
arcpy.management.MakeBuildingLayer(...)

#Package content
arcpy.management.CreateBuildingSceneLayerPackage(...)

#Custom function, delete older version of a building if it exists in AGOL
prep_arcgis_online(portal, title_temp)

#Share new building with specific tags and sharing level that let our ExB App to see the item 
arcpy.management.SharePackage(...)

The Georeferencing Hurdle

FITS recently received the Ellington Innovation Award from CFTA for a presentation outlining our work in this area. Georeferencing 773 BIM is a tedious process, but over years of effort, we accelerated this through a foundational ArcPy automation that became the basis for the production-level ETL processes outlined above. Due to the iterative work of the BIM and GIS teams on this, all BIM now contain real-world coordinates and require no additional georeferencing within GIS (only a .prj file is required).

An overview of Columbus Campus facing east with BIM and the university district buildings shown in contrast to the level of detail.

Upcoming Goals

So, how is this framework being used, and what comes next? Campus planning and operations remain core use cases. We have already completed asset collection projects related to classroom technology, informal furniture, public safety assets, art curation, and student/staff amenities. Moving forward, a major focus will be on tying more data together, building query tools, and indoor routing. Improvements to exterior shell extraction from BIM are also critical.

Spaces filtered to show restrooms and hallways in transparent buildings on a section of the Columbus Campus.

Parting Thoughts

Currently, the FITS GIS ecosystem is breaking down data silos. Built atop the bedrock of rich BIM and Space data, GIS acts as the spatial framework for our university.

Smart campus operations are often defined by their technical use cases. But I prefer to define them by the people doing the work. We provide a multidisciplinary perspective on all the physical areas that make up the university ecosystem.

We are… Architects, Geographers, and Space Data Managers.
For… Planners, Arborists, 911 Dispatchers, Professors, Students, Families, Patients, Medical Staff, Fans, and so many others.

If you have any questions or would like to further connect you can connect with me on LinkedIn or contact me at [email protected].