San Francisco International Airport (SFO) hosted 58 million passengers in 2018, is on a record-setting pace to peak 59 million passengers in 2019. These operations take place across a sprawling terminal complex of over 200 acres. This presentation will show how San Francisco International Airport uses FME to parse GIS data into a central database and ultimately into an IMDF model. The IMDF can then be consumed by multiple parties including mobile devices.
2. Sub-Surface Utilities
Data Model - UDM
Exterior Data Model - EDM Interior Data Model - IDM
MUNSYS Schema FAA AGIS (18B) Schema Hybrid SFO Custom Schema
SFO GIS
7. Campus Stats
Campus Area 3.54 square miles Landside (2,246 acres)
Buildings ≥ 1,000 sqft. (excluding MOC and USCG) 97
All Interior Area 16,431,155 sqft. (377 acres)
Terminal Area 5,049,203 sqft. (115 acres)
Terminal Public Area 2,358,000 (54 acres)
Employees 43,000
Passengers ~58,000,000
15. Output Destinations & Target Systems
One Map for All
GIS
3D
Model
CAD
PMBSABM
AVSEC
S/A
CAD
911
SFPD SFPD
Wayfi
nding
Apple
Maps
Airline
Maps
Google
Maps
AVM
Space
Util.
Meter
Lock-
Smith
Terminal
Systems
FIDS
ITT
WiFi
People
Flow
Lease
Structure
EMCS
CMMS
Gate
Travel
Times
FLYSFO
.com
Internal Apps
External Apps
Custom Apps and
Analysis
22. Database Model
• ERD of IMDF database design
featuring all feature types &
related tables
• Subset of Semantic Model
developed with an Airports
Council International (ACI)
working group
26. Level Processor
• Geometries reprojected into LL-WGS84
• Display Point extracted
• Joined with Buildings to get building_uuid
• level_uuid generated
• Metadata altered, columns renamed
27. Units
• IMDF defines Units as spaces
with floor to ceiling barriers
• Many of the spaces from the
source data do not fit that
criteria–will need to be
dissolved together
• Image to the left is color coded
by two attribute values:
– Blue represents PRIVATE
(nonpublic) spaces
– Other spaces color coded by
manual DISSOLVE CODE
– Purple spaces are PUBLIC and do
not have a dissolve code
28. Unit Processor
• Units are joined with corresponding level
• Appropriate spaces are dissolved together
• 4,569 PRIVATE spaces dissolved into 311
• 148 dissolve codes on 1,034 PUBLIC spaces
29. Before vs. After Transformation
The simplified geometries
reflects the level of
granularity desired for a
public-facing map
30. Openings
• Any break in the floor to ceiling barrier
• Doors
• Service counters
• Open entrances
• Many were drawn manually (indicated
by red lines)
• Much of the source data represented
doors by points (indicated in green),
while IMDF requires Openings to be LINE-
STRINGS
31. Idea from Dave Campanas (FME Product Expert):
• Line-strings are clipped out of Unit boundaries
by creating a buffer around the door point the
size of the door width and using the Clipper to
extract the Opening geometry
• Door widths had to be measured manually
Opening Processor
32. The created Openings are in
red. This trick allowed us to
leverage our existing data
instead of having to
recreate each door
34. Relationship Spec
• Was the most complicated feature
type to model
• Used to model correlations and
directions between certain Openings
and Units
• We used it for Escalators, Stairs,
Elevators, Moving Walkways, and
Unidirectional Doors in our model
35. Data Model
Relationship Data
Model
• Each relationship has a number of nodes
• Origin (1), Intermediary (0+), Destination (1)
• Each relationship has a category (Escalator,
Elevator, etc.)
• Each node has a type (Opening, Unit,
Escalator, Elevator, etc.)
37. Relationships Visualized
• Moving walkways in grey
• Escalators and stairs scattered throughout
• Elevators color coordinated by relationship
• Directed Openings represented by circles:
• Green for Origin
• Red for Destination
38. Last Custom Transformer: Clean Geometries
All geometries
sent through
Snapper; grouped
by level_uuid
39. Final Step: Database to IMDF
Normalized relational database consolidated into IMDF model
Agie will mention data – tiers
Tier 1: Exclusively maintained by SFO GIS
Tier 2: Co-Maintained by SFO GIS and some other stakeholder
Tier 3: Exclusively maintained by stakeholder and considered their data (mostly web apps and iPads)
Tier 4: Deliverable that will not be modified of maintained and will remain in it’s original state
Agie will mention data – tiers
Tier 1: Exclusively maintained by SFO GIS
Tier 2: Co-Maintained by SFO GIS and some other stakeholder
Tier 3: Exclusively maintained by stakeholder and considered their data (mostly web apps and iPads)
Tier 4: Deliverable that will not be modified of maintained and will remain in it’s original state
58 million PAX
43 thousand employees
Family and friends
Taxi + TNC