Form, Design & Simulation (3D + Impact) - CS-GY 6313 - Fall 2026
2026-11-06

Projection, normalization, classing

Urban flows, interactivity, linked views

Built environment, form, impact analysis
From flat maps to the three-dimensional form of the city
Explore how 3D visualization supports urban planning, architectural design, and impact simulation.
The Fundamental Question:
How do people mentally map a city?
Lynch’s Answer:
Not through abstract 2D maps, but through five types of elements that we experience in 3D.
A city is legible when those elements are easy to identify and connect.

Lynch, K. (1960). The Image of the City. MIT Press.


Same city, two methods: what residents said (left) and what they drew (right).
Lynch, K. (1960). The Image of the City. MIT Press. Figs. 35 and 36, p. 146.
1. Paths: channels of movement (dashed and solid lines)
2. Edges: linear breaks, such as the river and waterfront (hatched strips)
3. Districts: areas with common character (shaded and dotted areas)
4. Nodes: focal points and junctions (circles)
5. Landmarks: unique reference points (stars and triangles)
Most of these depend on height, mass or visibility, so they are fundamentally 3D concepts.

Lynch, K. (1960). The Image of the City. MIT Press. Fig. 3, p. 19: the visual form of Boston as seen in the field.
2D maps show:
But they miss:
3D visualization reveals:
These are essential for understanding how a city feels to inhabit.
By 2050: about 68% of the world’s population will live in cities
Making decisions about massive new developments:
How do we know the impact of these developments on neighborhoods, residents, and the urban fabric?

A 3D framework to support data-driven decision making in urban development
Ferreira, N., Lage, M., Doraiswamy, H., Vo, H., Wilson, L., Werner, H., Park, M., & Silva, C. (2015). Urbane: A 3D Framework to Support Data Driven Decision Making in Urban Development. IEEE VAST 2015.
The data management, impact analysis, and visual interface components of Urbane
Ferreira et al. (2015). Urbane, Fig. 2.
Polygon, grid, combined 2D+3D and road-network heat map layers over Manhattan
Ferreira et al. (2015). Urbane, Fig. 4.
Brushing and linking works in 3D, too!
Select buildings or a neighborhood in the map
↓
See their attributes in the data exploration view
Use case: “What are the characteristics of this neighborhood?”
Brush ranges on the parallel coordinates axes
↓
See matching buildings highlighted in the 3D map
Use case: “Show me buildings near a park with good subway access”
Each filter on the parallel coordinates plot updates which buildings are highlighted in the map.
Brushing a parallel coordinates plot (a, c) and the resulting buildings in the map (b, d)
Ferreira et al. (2015). Urbane, Fig. 5.
Compare the neighborhood to the rest of Manhattan, then filter down to candidate sites.
Parallel coordinates comparing neighborhoods (a), a filter on site properties (b), and the resulting candidate sites in 3D (c, d)
Ferreira et al. (2015). Urbane, Fig. 6.
The most powerful feature: “What-if” analysis
“How will a new building change its surroundings?”

Ferreira et al. (2015). Urbane, Fig. 3.

Ferreira et al. (2015). Urbane, Fig. 7.
Iterations: one or two
Iterations: as many as time allows
More informed decisions, and stakeholders who can see the trade-offs.
Why test one design? Explore the whole space of designs.
Doraiswamy, H., Wilson, L., Werner, H., Park, M., & Silva, C. (2015). Topology-based Catalogue Exploration Framework for Identifying View-Enhanced Tower Designs. ACM SIGGRAPH Asia (ACM TOG 34(6)).
Find tower designs that maximize view quality while staying practical:
A tower is generated from floor profiles and a few operations; the parameters define a design space.
Families of tower designs generated by varying parameters

For each location (x, y, z) and viewing direction θ, store what is visible
A view for any window is a constant-time lookup

A parallel coordinates view over design and view metrics lets the architect filter to the designs that matter.
Parallel coordinates for filtering the design catalogue

Selected design, refined profile, and program allocation by view quality
Exploration: minimal
Exploration: comprehensive
Visualization turns an enormous search into something a person can steer.
“3D? Is it the right way to go?” Same city, two representations: which question is each one good for?


Urbane’s solution: 3D for context, 2D for precision
3D for:
2D for:
Question: When is spatial encoding the right choice?
Question: How do we explore massive, dynamic flows?
Question: When is the third dimension essential?
When approaching any geographic visualization problem, ask:
The 2D lens (Week 8)
The temporal lens (Week 9)
The 3D lens (Week 10)
Every visualization makes choices.
Make yours deliberately, defensibly, and ethically.
