Amy X. Chen

Behind: fieldwork in Barcelona, 2025 · sketchbook, 2023–25 · New York street photography, 2023

Energy. Emerging tech. Abundance. Equity.

I make the physical, distributional consequences of urban systems legible to analysis.

Amy X. Chen

Harvard Graduate School of Design | Master in Urban Planning ’28
Columbia University | B.A. Economics–Mathematical Probability ’26

Now: Master in Urban Planning at Harvard GSD · research assistant, Healthy Places Design Lab · Google DeepMind travel grant to ACM SIGSPATIAL 2026 · writing The Algorithmic Commons

Start a conversation →What I found ↓Résumé ↓

Read this as
“A city is the place of availabilities. It is the place where a small boy, as he walks through it, may see something that will tell him what he wants to do his whole life.” — Louis Kahn
Map of the work

Hover a dot to read it. Click a project to open it, or an audience to tailor the page.

Six things I found

Each one links to the work behind it.

< ½

of the wait for a New York renewable-energy siting permit is covered by its 1-year legal deadline. The rest happens before the clock starts.

See the chart →40×

more of Chicago sits in the 100-year flood zone on a private flood model than on FEMA's map. Buyers now see the private score next to the price.

See the chart →52

of Chicago's 826 census tracts left for people to decide. AI set aside the other 774, which lose under every definition of flood risk.

Project 3.7 →116

more Citi Bike trips a day for each added mile of protected bike lane in New York.

Project 1.2 →61,228

Auckland buildings given heights from open LiDAR and rendered in a browser, no software license needed.

Project 1.1 →1st

place in Round 1 of the Google OASIS challenge at ACM SIGSPATIAL 2026, with an AI assistant for flood mitigation.

Project 3.7 →
Work

Thirteen projects, three questions

Open any card for the full project.

1.1Data systems Open-Data 3D City ModelingDigital surface and terrain model differencing, and an open-data contribution pipeline 1.2Data systems A Time Series Estimate of Protected Lane Returns on Citi Bike DemandWhat 50 miles a year of protected bike lanes does to ridership 1.3Data systems Good Friction, Bad Friction: Where AI Should Stop in PlanningA lightning talk on where AI belongs in planning decisions 1.4Data systems InfraScope: Data-Center Approval ScreeningWhen a town's rules are stored as data, you can see which change would flip a decision 2.1Regeneration & infrastructure Reimagine Manhattan: A Mathematical Modeling Approach for Superblock DevelopmentGraph models of the street network to find where superblocks would work 2.2Regeneration & infrastructure A Tale of Two Grids: Energy & ICT in Allston, BostonWho powers a neighborhood that Harvard and the data economy are rebuilding 3.1Climate risk & policy Climate Resiliency Modeling: Mobility Under Extreme WeatherThe June 2023 wildfire haze as a natural experiment 3.2Climate risk & policy Storytelling for Coastal Resilience: Staten Island North ShoreCommunity-led data on the MERC corridor 3.3Climate risk & policy Trucks, Fumes, and GroceriesEnvironmental justice in Mott Haven–Port Morris, South Bronx 3.4Climate risk & policy Spatial Regression: Rethinking Proximity and Risk in San FranciscoWho lives closest to flood risk, and what is it worth 3.5Curriculum and national sustainability reporting3.5Climate risk & policy Policy Case Studies on Sustainable CitiesCurriculum and national sustainability reporting 3.6Climate risk & policy From Reform to Crisis: Reconstructing the Public Narrative of NYC Public Housing, 1900–2026126 years of New York Times coverage, read as data 3.7Climate risk & policy Flood Smart Assistant: AI That Narrows, People Who ChooseGoogle OASIS 2026 Student Challenge, first place (Round 1), ACM SIGSPATIAL
Let's talk

Which friction in planning should we never automate away? And who gets to draw that line?

AI safety & governance

If some neighborhoods leave a thinner record, how do we keep these tools from widening the gap?

Evidence on where AI actually speeds up public decisions and where it shouldn't decide, plus the econometrics to measure its effects.

Where AI should stop · Flood assistant

Talk about this →
Planning & the public sector

Agencies are wary of the word AI but warm to a tool they can see working. How do we earn that trust?

Planning training at Harvard GSD and tools built to be adopted by agencies and community boards, not just published.

Permitting chart · Superblocks model

Talk about this →
Résumé

Experience, research & skills

Click a bar to jump to that role, filter by type, or pick a skill to see where it was used.

LinkedIn ↗GitHub ↗Email ↗
Show
202320242025202620272028Today
EducationExperienceResearchProjectsWhite notch = ongoing · dashes = planned
Advisory
Energy & Climate
Quantitative
Technical

Education2026–2028
Cambridge, MA

Master in Urban Planning (MUP), Concentration in Urban Analytics

Harvard University, Graduate School of Design

  • Coursework: Spatial Analysis, Data Science I, Land Use and Environmental Law, First-Year Urban Planning Studio
Data ScienceArcGIS ProPython
Education2022–2026
New York, NY

B.A. Economics, Mathematical Probability

Columbia University, Columbia College

  • GPA 3.889/4.00; Dean's List, 6 semesters
  • Coursework: Modern Analysis I, ODE, Probability Theory, Linear Algebra, Calculus I–III, Urban Economics, Econometrics, Financial Economics, Market Design, Political Economics, Advanced GIS, Urban Infrastructure Systems, Sustainable Urban Systems Engineering
EconometricsPanel Regression
ExperienceSept. 2026 – present

Research assistant

Harvard GSD, Healthy Places Design Lab

  • Analyze public datasets to find built-environment proxies for heat vulnerability, and turn the findings into decision-support frameworks for planners and public-health stakeholders.
Climate AdaptationData SciencePolicy Research
ExperienceSept. 2026 – present

Peer mapping assistant

Harvard GSD, Center for Geographic Analysis

  • Technical GIS consultations on spatial data, ArcGIS Pro, QGIS, cartography and geospatial visualization for planning and design projects.
ArcGIS ProQGISStakeholder Communication
ExperienceMay 2026 – present

Product development intern

Block Party (civic tech startup)

  • Turned a 22-year archive of 4,589 Manhattan Community Board 3 resolutions into a quality-controlled geospatial dataset, with automated extraction, multi-geocoder validation and confidence scoring to map land-use, capital and service priorities; the pipeline scaled to all 59 NYC community boards.
  • Presented the product to Harvard Data-Smart City Solutions fellows, industry professionals and NYC public agencies.
PythonData ScienceStakeholder CommunicationSQL
ExperienceJul. – Aug. 2025
Madrid

Business innovation & strategy intern

Mormedi (strategic design and innovation consultancy)

  • Comparative infrastructure-feasibility analysis of international aerial tramway systems (operating costs, ridership, first/last-mile integration) to support corridor and service recommendations for Madrid's regional transit agency.
  • Market sizing and demand analysis for Seattle eVTOL networks, with phased deployment scenarios reflecting regulatory, network and siting constraints.
Feasibility AnalysisMarket SizingScenario PlanningInfrastructure StrategyInfrastructure SitingExcelPowerPoint
Experience2023–2025

Research assistant & curriculum co-developer

Columbia Climate School, Sustainability Policy & Management

  • Benchmarked municipal climate policies and sustainability indicators across Lima and Copenhagen; research synthesized for reports presented at COP29 and COP30.
  • Developed a 200-page sustainable-cities module on municipal waste infrastructure (material recovery, landfill diversion, sanitation, implementation economics); funded by Tencent.
Policy ResearchMunicipal SustainabilityDecarbonization
See project 3.5 ↓
Research2026 – present

Sole author & presenter: From Reform to Crisis, sentiment on public housing in the Lower East Side, 1900–2026

2026 NYU Urban Systems Symposium

  • Built a 10,118-article corpus from the New York Times Archive API (1900–2026); VADER sentiment scoring and Bai–Perron structural-break detection to find shifts in public discourse on public housing.
  • Presented at the 2026 NYU Urban Systems Symposium; manuscript published in the conference proceedings.
PythonEconometricsData SciencePolicy Research
Paper ↗See project 3.6 ↓
ResearchMay 2025 – present

Contributing author: Multi-Modal Transit Resilience Under Climate Disruption

Columbia Urban Systems Engineering Lab

  • Built origin–destination flow matrices and used NYC wildfire-haze AQI spikes as a natural experiment to estimate cross-modal (bikeshare, bus, taxi, subway) demand elasticity and neighborhood rerouting vulnerability.
EconometricsPanel RegressionPythonClimate AdaptationData Science
Paper ↗See project 3.1 ↓
ResearchFeb. – May 2025

Co-author: Coastal Flood Exposure & Property Value Analysis, San Francisco

Columbia, Advanced GIS

  • Modeled 1D–3D terrain-based flood exposure and tested its relationship with property values and socioeconomic indicators; designed and presented a 20-page planning report.
ArcGIS ProIllustratorInDesignCoastal ResilienceEconometrics
Paper ↗See project 3.4 ↓
Research2024

Summer fellow: Superblock Feasibility & Street Network Topology, Manhattan

Columbia Department of Mathematics

  • Modeled Manhattan as a graph of 5,000+ intersections and road segments with OpenStreetMap features and distances to build a Superblock Suitability Score, and simulated a phased implementation.
PythonMulti-Criteria AnalysisScenario PlanningData Science
Paper ↗See project 2.1 ↓
ProjectsAug. 2026 – present

Google OASIS 2026 Student Challenge, First Place (Round 1)

ACM SIGSPATIAL 2026

  • Co-authored A Dimension-Preserving AI Assistant for Flood Mitigation: Multi-Scale Pareto Analysis in Chicago, a decision-support framework for comparing interventions across spatial scales.
  • Awarded a $1,000 Google DeepMind Student Travel Grant to represent the team at ACM SIGSPATIAL 2026.
Climate AdaptationMulti-Criteria AnalysisPythonData Science
See project 3.7 ↓

Languages: English, Mandarin (native)

Part 4

Data Gallery

Charts and interactive tools on how data and AI can make real estate, infrastructure and planning decisions faster and better grounded. Every number is sourced, with links and page numbers.

Charts

4.12026-10 · Climate risk · Real estate
A private flood model puts 40 times more of Chicago in the 100-year flood zone than FEMA

A private flood model puts 40 times more of Chicago in the 100-year flood zone than FEMA

FEMA's map puts 0.3% of Chicago properties in the 100-year flood zone; First Street's model, which counts heavy rain, puts 12.8% there. Buyers now see the private score next to the price.

LinkedIn post: Climate risk is priced at the search

Rebuild this chart: notebook ↗
4.22026-10 · Permitting · Infrastructure
New York's 1-year deadline covers less than half the wait for a siting permit

New York's 1-year deadline covers less than half the wait for a siting permit

New York met its 1-year legal deadline on every renewable siting permit, but the wait before the deadline starts (getting an application declared complete) averaged 1,094 days in 2020–23.

LinkedIn post: AI won't fix permitting on its own

Rebuild this chart: notebook ↗
4.32026-10 · Permitting · Infrastructure
Same approval, months or days

Same approval, months or days

By-right housing in LA was permitted 60–248 days faster; automated solar permits take 0 days vs 7; New York's 365-day clock sits inside a 1,333-day process.

Talk: Making Progress

Rebuild this chart: notebook ↗
4.42026-10 · Climate risk · Real estate
Who counts as most at flood risk depends on the definition

Who counts as most at flood risk depends on the definition

FEMA maps 0.3% of Chicago properties in the 100-year zone; First Street maps 12.8%. Four reasonable definitions of risk agree on almost no census tracts.

Talk: Making Progress

Interactive

4.52026-10 · Governance
What Planners Decide

What Planners Decide

Explorable bubble charts of 19,000+ decisions at NYC City Planning, Manhattan community boards, Boston and SF, with a real example for each topic.

Open project ↗
4.62026-04 · Climate risk · Maps
The Bronx River: A Story of Return

The Bronx River: A Story of Return

Two decades of restoration on an urban river: herring, dams, and the 1.4 million people who share the watershed.

Open project ↗
Appendix

Spatial Representation & Field Observation

Drawing, photography and writing: the habits of looking closely that the data work depends on.

M1
Observational Drawing

Observational Drawing

Drawing from life at Fontainebleau, Ronchamp, the Gilder Center and Central Park, 2023–2025.

M2
Urban Journalism: Untapped New York

Urban Journalism: Untapped New York

Editorial and photography internship, 2023: articles on Coney Island Hospital's oldest building, the NYBG summer exhibit and eco-friendly sites in Manhattan.

M3
The Algorithmic Commons & Other Essays

The Algorithmic Commons & Other Essays

Essays in Urban AI, the Columbia Daily Spectator, Barnard Columbia Urban Review and the NZ Institute of Architects, plus a Substack, The Algorithmic Commons.