Imagine a world where we have digitized information about every building in the US that is freely available. What could be the impact? We’ve just launched the Open City Model to find out.
The lack of easily available digital data about building stock in the US is stifling local efforts to make plans to fight climate change. Open City Model has been designed and launched to solve this problem and is tackling it head on by releasing a completely open dataset providing 3D models of 125M buildings in the US conforming to the OGC standard CityGMLurban model.
Open City Model uses a modern data pipeline built entirely in the cloud using 100% open source tools. Our initial data release representing over 200GB of open geospatial data was produced for less than $300 total computing costs.
2. About me!
● Born and raised in California
● SCU grad - software engineer
● Startup guy until 2016
● Co-founded BuildZero to fight climate change
● New to GIS!
BuildZero.Org
3. The 20-minute Tour
What is Open City Model?
How is it created and maintained?
What are the future plans?
Questions
10m
2m
8m
5m
4. The 18-minute Tour
What is Open City Model?
How is it created and maintained?
What are the future plans?
Questions
10m
2m
8m
5m
6. It all began when ...
Start to design some cool
software
BuildZero.org begins
7. It all began when ...
Start to design some cool
software
Search for data about
buildings
BuildZero.org begins
8. It all began when ...
Start to design some cool
software
Search for data about
buildings
Ugh. Finding building data is
hard
BuildZero.org begins
9. It all began when ...
Start to design some cool
software
Search for data about
buildings
Open City Model!
Ugh. Finding building data is
hard
BuildZero.org begins
11. Open City Model
3D geometries, building identifiers, and common attributes
CityGML & CityJSON
01
02
03
04
05
125M buildings, 5k files, 250GB uncompressed
Full USA coverage
Hosted on S3, free for download
AWS Public Dataset
github.com/opencitymodel/opencitymodel
Open-source data pipeline
Help us with code, data sourcing, documentation, and more!
Community supported
CityGML data for every building in the US
12. Why CityGML?
● OGC Standard
● Stable ongoing development
● Robust spec for 3D geometries & more
● Extensible as part of its core design
● Multiple data formats (GML, JSON)
16. Prep Enhance Merge Format Publish
The OCM Data Pipeline
100% open source code on github run via AWS using a serverless approach.
One engineer (6m part-time), 2 releases, 3 source datasets, full US coverage, $300 compute cost.
Custom AWS Batch
Docker
Javascript
EMR + Spark
Scala
AWS Batch
Docker
Java/citygml4j
The pipeline is composed of independent jobs, typically defined via Docker images, which can easily be
submitted to run on a cloud compute platform without maintaining any servers.
21. Future plans
Smarter entity resolution
Improved machine learning
Algorithms
Open Street Maps, local govs,
and more
Data sources
Open
City
Model
22. Future plans
Smarter entity resolution
Improved machine learning
Algorithms
Open Street Maps, local govs,
and more
Data sources
LOD1 -> LOD2/3 -> LOD4
Other city objects
New Features
Open
City
Model
23. Future plans
Smarter entity resolution
Improved machine learning
Algorithms
Open Street Maps, local govs,
and more
Data sources
Get more people involved!
Community
LOD1 -> LOD2/3 -> LOD4
Other city objects
New Features
Open
City
Model
24. The 18-minute Tour
What is Open City Model?
How is it created and maintained?
What are the future plans?
Questions
10m
2m
8m
5m