The Origins of Gothic at the Church of Saint-Denis

Written with Stephen Murray, medieval historian at Columbia University
This is my Columbia University senior thesis in the History & Theory of Architecture. This work expands on my films about Amiens Cathedral, published here.

Abstract

Around the year 1140 CE, a new style of architecture and way of thinking about how to construct buildings developed in Northern France. This way of building soon spread across Europe, seeding cathedrals, monasteries, abbeys, and churches wherever masons traveled. Centuries later – long after masons ceased building in this style – Renaissance architectural theorists began calling this style the “Gothic.”
The one church traditionally associated with this 1140s stylistic shift from the earlier Romanesque style to the newer Gothic style is a small building just north of Paris: the Abbey Church of S-Denis. However, although the popular narrative of architectural history assumes this building to be the world’s first Gothic building, little structural evidence to this effect survives. This thesis follows two strains of inquiry: 1) why this church is associated with the origins of Gothic and 2) how surviving fragments of the 1140s S-Denis fail to support claims of the structure’s primacy.
Why does this matter? S-Denis reveals a tendency to tell history – particularly architectural history – in terms of individual structures when, in fact, the origins of the Gothic style might be more complex. By abandoning a Paris and S-Denis centric origins story, we might be able to better appreciate the diverse array of local sources from which medieval masons found inspiration to build.

.

Read the thesis online

Opens as PDF in new window

.

Strangely enough, despite the accepted fact that S-Denis’ architecture was significantly rebuilt, numerous sources continue to assume this church to be the first. Copied below is a quote from S-Denis’ official website:

.

The birth of Gothic art. The church, designed by Abbot Suger, kings’ advisor from 1135 to 1144, was completed in the 13th-century during the reign of Saint Louis. A major work of Gothic art, this church was the first to place a great importance on light, a symbol of divinity, in religious architecture.

.

Or this quote from leading German medievalist Dieter Kimpel:

.

Suger, abbot of the most important of all the royal abbeys, that of Saint-Denis, and sponsor of the western part and the sanctuary of the abbey church, works considered rightly as a milestone in the history of the birth of Gothic architecture, left us a detailed account of his activity as abbot.

.

24 Hours in the London Underground

Audio effect: Heartbeat from Freesound

.

Through analyzing 25,440 data points collected from 265 stations, this animation visualizes commuting patterns in the London Underground over two weeks in 2010.
Each colored dot is one underground station. The dots pulsate larger and smaller in mathematical proportion to the number of riders passing through. Big dots for busy stations. Small dots for less busy stations.
Dot color represents the lines serving each station. White dots are for stations where three or more lines intersect. Each dot pulsates twice in a day: Once during the morning commute; and again during the evening commute.
By syncing the audio volume with the density of riders and the background color with the time of day, the animation becomes acoustically legible. The audio volume rises and falls to mirror the growth and contraction of each colored dot during the daily commute.

.

.

The rhythmic pulsing of commuters is analogous to the breathing human body. The passage of red blood cells from the lungs to the organs is analogous to the movement of people to and from the city’s own heart: the downtown commercial district. This analogy of human form to city plan is a longstanding theme in urban studies.
See my film about commuting patterns in the NYC subway.

.

The Data

.

.

Method

No single data set could capture the complexity of a metropolis like London. This animation is based off of open-access data collected in November 2010. According to Transport for London: “Passenger counts collect information about passenger numbers entering and exiting London Underground stations, largely based on the Underground ticketing system gate data.” Excluding London Overground, the Docklands Light Railways, National Rail, and other transport providers, there are 265 London Underground stations surveyed. For data collection purposes, stations where two or more lines intersect are counted as a single data entry. This is to avoid double-counting a single passenger who is just transferring trains in one station en route to their final destination.

Every fifteen minutes, the numbers of passengers entering the system are tallied. This yields 96 time intervals per day (4 x 24). Multiplying the number of time intervals (96) by the number of stations (265), we get the number of data points represented in this animation: 25,440. Each station was assigned:

  • A location on the map of latitude and longitude
  • A color according to the lines extant in 2010: Bakerloo, Central, Circle, District, Hammersmith & City, Jubilee, Metropolitan, Northern, Piccadilly, Victoria, Waterloo & City.
  • A circle scaled to reflect the number of passengers moving through. Stations range in business from a few hundred passengers to over 100,000 per day.
  • A time of day: each 15-minute interval becomes one image in this film. Overlaying these 96 “snapshots” of commuter movement creates  a time-lapse animation. Thus, a single day with 25,440 data points is compressed into a mere 8 seconds.

Sources

Station Coordinates: Chris Bell. “London Stations.” doogal.co.uk (link)
Ridership Statistics: Transport for London. “Our Open Data.” (link)
Click on the section “Network Statistics” to view “London Underground passenger counts data.”

.

Powered by TfL Open Data. Contains OS data© Crown copyright and database rights 2016.

Railroad commuting patterns in New Jersey

View my data visualizations of New Jersey’s suburban growth here.
Created with data from NJ Transit on weekday and weekend rail ridership.
Or download my data from Tableau Public.
NJ Transit carries over 90,000 commuters per day to and from New York Penn Station, the busiest rail station in the Western Hemisphere. The construction of this rail network in the nineteenth and early twentieth centuries was focused around New York City. Like spokes on a wheel, these rail lines radiate from the urban center.
Hover over stations to view statistics. Dot color corresponds to train line. White dots are for stations where multiple lines intersect. Dot size corresponds to number of riders per day: Large dots for busy stations and small dots for less busy stations. For each station, the average number of daily riders is listed.

.

.

The map above shows weekday ridership patterns. Movement is centered around the employment hubs of Newark and New York Penn Station. The next two busiest stations are Secaucus Junction and Hoboken, but these two stations are not destinations. Instead, they are transfer points for commuters en route to New York City. Commuters collected from stations on the Pascack Valley, Bergen County, and Main Line are almost all headed to New York City, but they must transfer at Secaucus (to another NJT train) or at Hoboken (to PATH / Hudson River ferries).

.

.

This map shows Sunday ridership. On average, stations are 66% to 75% less busy on weekends. The thirteen stations along the Montclair-Boonton Line – between Bay Street and Denville – are also closed on weekends because ridership is so low. However, the only line that is almost as busy on weekends as it is on weekdays is the Atlantic City Line. This is likely because trains on this line serve weekend tourists to the New Jersey Shore and Atlantic City casinos.

.

.

Notice the large difference between the first four stations and all others listed. Keep in mind that a lot of this data double-counts a single passenger. For instance, someone riding from their home to work will be counted once in the morning, and again in the evening.

.

Writing Here Is New York in 1949, American writer E.B. White has this to say about suburban commuters:

.

“The commuter is the queerest bird of all. The suburb he inhabits […] is a mere roost where he comes at day’s end to go to sleep. Except in rare cases, the man who lives in Mamaroneck or Little New or Teaneck, and works in New York, discovers nothing much about the city except the time of arrival and departure of trains and buses, and the path to a quick lunch. […] About 400,000 men and women come charging onto the Island each week-day morning, out of the mouths of tubes and tunnels. […] The commuter dies with tremendous mileage to his credit, but he is no rover. […] The Long Island Rail Road alone carried forty million commuters last year, but many of them were the same fellow retracing his steps.” (p.18-21)

Northeast Corridor railroad time-lapse

Audio effects from Freesound; music is Metamorphosis by Philip Glass

.

The Northeast Corridor is the busiest passenger railroad in North America. This drone flight follows a high-speed Acela train making this 456 mile journey from Washington D.C. to Boston via Baltimore, Wilmington, Philadelphia, Trenton, Newark, New York City, Stamford, New Haven, and Providence.
This animation was created from Google Earth satellite imagery. I traced the Northeast Corridor route onto the ground, and I then programmed the computer to follow this route. I then added the inset map, sound effects, and clock in post-production.
The above animation is condensed. View the full and uncut 28 minute flight here.

Geography of Marijuana Arrests

Update March 2021: Marijuana is now legal in NY state.

 

.

The New York Police Department (NYPD) made 102,992 arrests in 2017 for the possession, sale, and/or use of marijuana. 1 While only 25.5% of New Yorkers are Black, 67.5% of marijuana arrests are of Blacks. Similarly, 90% marijuana arrests are male, even though only 65% marijuana users are male. 2 Males more than females and Blacks more than others are arrested for marijuana in disproportionate numbers.

.

Race
Percentage of New Yorkers who identify as this race 3
Percentage of marijuana arrests of individuals belonging to this race
White
44.0%
11.2%
Black
25.5%
67.5%
Asian/Pacific Islander
12.8%
4.2%
Other
17.7%
17.1%

.

 

.

2017 data

.

Click table to view in detail

NYPD marijuana arrests are disproportionately of Black males between the ages of 18 and 44 from low-income communities, even though this demographic represents less than 10% of the city’s population. Why should this matter? Arresting individuals for using a relatively harmless and non-addictive drug is expensive for taxpayers. According to the Drug Policy Alliance, the city spends $75 million on marijuana arrests and prosecution per year. 4 This is money that could have gone to education, parks, and community programs. Marijuana policy targets our country’s poorest people of color.
The common argument, and the grounds on which marijuana was initially made illegal, is that marijuana is a “gateway drug.” Marijuana supposedly introduces and later encourages individuals to experiment with more dangerous and addictive substances. Whether or not this is true, the arrest and punishment of individuals for marijuana may incur the equal risk of becoming a “gateway crime” to the legal system. With a prison record from a marijuana arrest, a person of color may have more difficulty finding employment and re-entering society – ironically pushing them to desperation and possibly new and greater crimes than their initial arrest.

.

.

View this pie chart in more detail.

.

Below are three maps of neighborhood “hotspots” for marijuana arrests. The income of every block is indicated on a red to green color scale from low to high income. The population of Latinos and Blacks per square mile is also indicated; unsurprisingly, these groups cluster in low-income neighborhoods. On this base map is the geo-referenced address of every arrest for marijuana possession or sale from 2013 to 2017.
Marijuana arrests tend to happen in low-income neighborhoods. For instance, Manhattan’s 96th Street represents an income divide between the wealthy Upper East Side and the comparatively poorer Harlem. Drawing a “thin blue line” down 96th Street, we also identify an unspoken policing boundary. Marijuana arrests are significantly less likely to happen in the majority white neighborhood south of 96th than in the majority black neighborhood north, even though both neighborhoods are of comparable population density and likely comparable rates of marijuana use. According to the UCLA: “Despite roughly equal usage rates, Blacks are 3.73 times more likely than whites to be arrested for marijuana.” 5 Similarly, the wealthy and majority white neighborhood of Riverdale in the Bronx has few arrests in comparison to the poorer and majority black West Bronx, even though these two neighborhoods are less than mile apart.

 

.

.

Research Method

.

Note that on the above map, there are numerous low-income neighborhoods without any drug arrests. This is largely because these areas have little to no population, such as Central Park or LaGuardia Airport. Controlling for population density, marijuana arrests still target communities of color.
This project was assembled from public data. I downloaded anonymized microdata on the race, crime, gender, and approximate age of every individual arrested by NYPD, as well as the address where this individual was arrested. Of the approximately 1.7 million arrests in this data set, I filtered out the marijuana crimes. The colored basemap indicating per capita income and race by city block is extracted from Tableau Public, the mapping software I use. The infographics presented above can be explored or downloaded here. Arrest data is from NYC Open Data here.

.

Endnotes

  1. Marijuana arrests represent 5.98% of all NYPD arrests in 2017.
  2. From “Statista,” accessed 15 January 2019, link.
  3. From the United States Census Bureau, 2010 statistics on NYC demographics, link to report, link to database.
  4. From the Drug Policy Alliance, accessed 15 January 2019, link to press release, link to report.
  5. From the American Civil Liberties Union, accessed 18 January 2019, link to article.

New York City Subway Ridership

Created with data from the MTA.
Published by Gothamist on 22 January 2019.
Related: my data visualization of London Underground commuting patterns.

.

The visual language of data addresses a deeper need to humanize and soften the concrete jungle.

.

Sounds of breathingheartbeat, and subway from Freesound

.

In this animation based on subway ridership statistics by station:
● Dots are color-coded according to the subway lines they serve.
● White dots are for junctions between two or more lines of different color.
● Dot size corresponds to the number of riders entering each station within a 24 hour period.
● Larger dots are for busier stations. Smaller dots are for less busy stations.
Movements through the New York City subway are analogous to rhythmic breathing.
People often describe cities in relation to the human body. Major roads are called “arteries” in reference to blood flow. The sewers are the city’s “bowels” in reference to our own digestive systems. Central Park is the city’s “lungs.” At various times in history, key industries like garments and finance were described as the “backbone” of New York’s economy. Although cities are complex organisms, wordplay makes the giant metropolis somehow more human and familiar.
The 424 subway stations and 665 miles of track are analogous to the human circulatory system. Every weekday pre-coronavirus, the subway carried 5.4 million people, mostly commuters. This daily commute is ordered, structured, and rhythmic – as Manhattan’s population swells during the daily commute and then contracts by night. Each passenger symbolizes the movement of a single red blood cell. With each paycheck, the oxygen of capitalism flows from the heart of Manhattan to the cellular homes in the outer boroughs.
Commuting patterns mirror the rhythmic expansion and contraction of the human body while breathing. By contrasting weekday and weekend ridership patterns, we detect the city’s respiratory system.

.

.

Interactive Map

.

Research Method

In this video lecture, I walk you through how I manipulated MTA and NYC open data
to create this animation.

.

The Metropolitan Transit Authority (MTA) publishes statistics on weekday and weekend (Saturday + Sunday) ridership for all 424 stations. These statistics, updated yearly, are public and can be analyzed to track trends in urban growth. I downloaded the MTA data and assigned each station a geographical coordinate (latitude + longitude) so that the data points would appear at their corresponding map locations.

I have a love-hate relationship with the New York City subway. At rush hour, it is crowded, hot, and slow. From years of riding its squeaky trains, it’s given me a ringing tinnitus sound in my ear. Despite its flaws, the subway is one of the few urban spaces where all social classes and ethnicities mix, where their separate lives are momentarily shared. Rich or poor, everyone rides the subway. I hope this animation renews appreciation for this engineering and the people behind it.

.

Sources

Where in the world is modernism?

What if the nationality of every artist represented in the Museum of Modern Art’s collections were mapped to illustrate the museum’s evolving geographic diversity through time? Watch the data visualization below of 121,823 works at MoMA.

.

.

Introduction

“The Museum of Modern Art (MoMA) acquired its first artworks in 1929, the year it was established. Today, the Museum’s evolving collection contains almost 200,000 works from around the world spanning the last 150 years. The collection includes an ever-expanding range of visual expression, including painting, sculpture, printmaking, drawing, photography, architecture, design, film, and media and performance art.
“MoMA is committed to helping everyone understand, enjoy, and use our collection. The Museum’s website features 79,870 artworks from 26,215 artists. This research dataset contains 135,804 records, representing all of the works that have been accessioned into MoMA’s collection and cataloged in our database. It includes basic metadata for each work, including title, artist, date made, medium, dimensions, and date acquired by the Museum. Some of these records have incomplete information and are noted as ‘not Curator Approved.’
“The Artists dataset contains 15,757 records, representing all the artists who have work in MoMA’s collection and have been cataloged in our database. It includes basic metadata for each artist, including name, nationality, gender, birth year, death year, Wiki QID, and Getty ULAN ID.” – from MoMA’s website.
I downloaded this dataset and dissected it with this question in mind:
What trends might this dataset reveal about the history of curating and the growth of a museum’s collections?
In the three interactive features below, hover over the graphs to explore the data in depth.

.

1. Geographic and Gender Diversity

This map visualizes the nationalities of ~15,757 artists whose work is displayed at MoMA. There are 121,823 data entries displayed below. The data can be browsed by year or by department. This illustrates the evolving geographic breadth of collections. Beginning in the 1930s, over 80% of artworks were from the four key countries of the US, UK, France, and Germany. Beginning the 1960s, the museum acquired some of its first works from Latin America and Japan. Post-1991, the museum acquired the bulk of its collections from Russia and China. Recent years have also seen a slight growth in collections of African art
An important distinction: This map does not show where each artwork was made. Rather, it shows where each artist is from. Nationality and national identity are, depending on the artist, an important influence shaping the unique perspective artists bring to their work.

.

.

The bar chart below shows the gender distribution of artworks by date. On the horizontal axis: the date acquired. On the vertical axis: the number of artworks acquired in this year. Each bar is divided into three colors: Blue for artwork by a male artist. Pink for art by a female artist. Grey for art where the gender of the artist is not known.
This data can be explored by year and by department. Across departments, male artists represent the clear majority. The departments with the greatest number of works by female artists: Photography and Drawings. The department with the least female representation: Prints & Illustrated Books. The department with greatest number of works where the artists’ gender is unknown: Architecture & Design. However, across departments, the representation of female artists has slightly increased over the past few decades from around zero to somewhere around 20%.

.

.

2. Do newer acquisitions tend to be smaller?

The two graphs below plot the relationship between year produced, year acquired by MoMA, and the dimensions of each artwork (width in cm). I’ve plotted 12,250 points. They are color coded with the same blue, pink, and grey system as the previous chart.
In the first graph, we see that newly produced paintings are becoming progressively larger. In 1929, the year of MoMA’s founding, the width of the average painting being produced was less than 100cm. Today, the average width of newly produced paintings is around 400cm – and is steadily increasing.

.

.

In the second graph, we see that MoMA’s new acquisitions are becoming progressively smaller, even though newly produced artworks are larger than before. In 1929, the average width of a new acquisition was over 300 cm. Today, the width is less than 150cm.

.

.

In other words, while artists seem to be working in ever larger dimensions, MoMA seems to be acquiring ever smaller paintings from these artists. Have the growing costs of buying and storing art priced MoMA out of acquiring larger artworks? What is the relationship between size and the decision whether or not to acquire a work?

.

3. Is the scope and definition of modernism expanding?

The challenge facing any museum dedicated to modern art is keeping up-to-date. Modern art is constantly being produced. Like any leading museum, MoMA is:
  • growing its collection of newly-produced contemporary works
  • while also enhancing its collection of older works
  • and expanding the geographic and national representations of artists and artworks
The graph below compares the relationship between production year and acquisition year for 7,797 data entires. Dot size indicates the size of the acquisition (i.e. number of pages or number of paintings from said artist). The red trend line indicates the linear relationship between when a work was produced (vertical axis) and when it was acquired by MoMA (horizontal axis). The vertical gap between the trend line and the upper reaches of the graph indicates the time elapsed between when the work was produced and when it was acquired. With time, the number of years elapsed between production and acquisition has grown.
In 1929, most new acquisitions were produced in the 1920s – modernism was a new movement and a new idea. Today, new acquisitions range in date from the early nineteenth century through present day. The temporal definition of modernism is growing, with origins that stretch ever further back in time.

.

.

Modernism is not geographically restricted. With globalization and the march of capitalism, the world is becoming more modern and interconnected. As new regions adopt modern technology, materials, and ideas, the character of art and artists will change. Cultural institutions, particularly modern art museums, are positioned to curate these global trends through the kinds of works they acquire and display. However, the kinds of stories museums and curators can tell are limited by the size and diversity of the collections available.

.

Related Data Projects

.

Sources

Download MoMA’s data from GitHub. The analysis above reflects this dataset as of 17 October 2018. New entries after this date are not included as these infographics are not updated in real-time.

Download my analysis of this data and the infographics above from Tableau Public.

A History of Historic Preservation in New York City

Data analysis of NYC landmarks since 1965 reveals trends and biases in the landmarks preservation movement.

Developed with urban historian Kenneth Jackson at Columbia University’s Department of History

.

.

A visual history of landmarks preservation in NYC. Data from NYC Open Data. Music from Freesound.

.

Introduction

There is ongoing debate between in NYC between developers seeking to rebuild the city in the image of global capitalism and preservationists seeking to slow the rate of change and protect the appearance of the city’s many and distinct neighborhoods. Several factors drive historic preservation: fear of losing heritage; fear of change; historians, public servants, and well-intentioned activists in the spirit of Jane Jacobs. This debate has played out every year since 1965 through the hundreds of structures that are added to (or rejected from) the Landmarks Preservation Commission’s running list of landmarks (LPC). Once added, landmarked buildings cannot be modified without first seeking approval from the city. Landmarks preservation is contentious for developers because the protections of preservation law are permanent and affect all current and future owners. Preservation law further restricts significant rebuilding, even if demolition and rebuilding are lucrative for the property owner.
Historians decide the future of the city’s built environment. The sites they preserve will become the architectural lens through which future generations will appreciate the past. The sites they approve for demolition will be lost to history. Preservation is a response to larger historical questions: Which aspects of the past are worth preserving? How should the city balance the economic need for development with the cultural need for history?
This paper will assess the landscape of historic preservation through analysis of publicly-available landmark records from NYC Open Data. We identified two datasets, both containing ~130,000 spreadsheet entries for every single LPC listing from 1965 to 2019. The first dataset is titled “Individual Landmarks” 1 and includes the structure’s address, lot-size, and date landmarked. The second dataset is titled “LPC Individual Landmark and Historic District Building Database” 2  and includes the construction date, original use, style, and address of all structures. We downloaded both datasets as .csv files, imported them into a visualization software called Tableau, merged them into a single map, and then analyzed the data. The results of inform the conclusions presented here. This analysis is broken into four case studies:
  1. Distribution of Landmarks over the Five Boroughs
    Assesses where landmarks preservation is densest or least dense by neighborhood.
  2. Contextual Preservation?
    Analyzes how protecting a landmark limits redevelopment of neighboring properties of less aesthetic value
  3. How does the preservation movement reflect economic patterns?
    – Factor affecting the preservation of city-owned structures
    – Factors affecting the preservation of residential structures
    – Relationship between preservation and gentrification?
  4. Keeping up to pace?
    Questions the degree to which landmarks preservation succeeds in protecting recently-built landmarks
From this data, hidden trends and biases in historic preservation become visible. Firstly, we identify a higher-density of landmarks in certain (and usually higher income) neighborhoods. Secondly, we identify a marked preference among historians for protecting structures pre-1945. (Is there so little in the city’s recent architectural history that is worth preserving?) And thirdly, our analysis hints at the strength of market forces and developers in shaping the scope and definition of preservation.

.

.

Study One:

Distribution of Landmarks over the Five Boroughs

.

The tree map below shows the distribution of all 128,594 landmarks across the city. This includes both buildings and non-buildings like street lamps, parks, statues, and bridges. Each rectangle is scaled to reflect the number of landmarks within that borough’s historic district – the larger the box, the more buildings. The largest rectangle for each borough represents the number of individual landmarks that fall outside any historic district. Boxes are grouped and colored by borough: Manhattan, Brooklyn, Queens, Bronx, and Staten Island.

.

125,594 records above

.

Several trends are visible. For instance, Manhattan, with a population 19.3% of the citywide total, 3 has 30.46% of the landmarks. By comparison, Staten Island, with 5.55% of the population, has 16.24% of landmarks, which is the greatest number of landmarks relative to the smallest population. By contrast, the Bronx with 17.06% of the population has only 5.36% of landmarks, which is the least number of landmarks relative to population size and density.
Given that the Bronx’s land area (42.47 mi²) is comparable to Staten Island (58.69 mi²), and given that their histories are both rich, then does the Bronx objectively have fewer landmarks worth preserving? Or do preservation trends follow patterns of economics and race – with wealthy neighborhoods having stronger legal and political leverage to preserve their built environment?

.

Manhattan Brooklyn Queens Bronx Staten Island

% of NYC population in this borough

(8.623 million total)

19.30% 30.72 27.36 17.06 5.55
% of NYC landmarks in this borough 30.46% 25.65 21.98 5.36 16.24
Surface area 22.82 miles2 69.50 108.10 42.47 58.69

.

Historic preservation does not operate off of a tabula rasa with objective standards and processes. There are spatial patterns to urban growth and income inequality; privilege (or the lack of privilege) is concentrated in specific neighborhoods. The geography of historic preservation follows similar patterns.

.

128,212 records above

.

Study Two:

Contextual Preservation?

.

A common criticism is that preservation stifles economic development. Preservation prevents demolishing and replacing older structures with larger and more profitable new ones. A lengthy (and expensive) approvals process is also required to modify old buildings. A city committee reviews applications and suggests revisions to ensure that new development is “contextually” respectful of its neighbors and/or preserves as much of the historic building’s fabric as possible. 4
Developers often claim historic preservation discourages development and reduces profits. Our data does not support this claim. Developers claim that protecting one building can limit the redevelopment of neighboring buildings. This criticism applies to vacant parcels within historic districts. This critique also applies to non-historic and non-landmarked buildings that fall within historic districts, but whose redevelopment might weaken neighboring landmarks. Construction vibrations and foundation vibrations on non-historic properties often destabilize and damage nearby landmarks, which is something developers need to address when seeking approvals from the city.
Within the city’s unequal fabric with pockets of concentrated wealth, poverty, and history, there are three general categories of protected buildings.
Firstly, there are individual landmarks, such as bridges, train stations, statues, and street furniture. While aesthetically and historically important, these sites are stand-alone pieces. New development can occur nearby with few restrictions. Historical review is not required for neighboring properties; only construction permits are needed. The case for protecting individual landmarks is strong; the nomination was written and approved on a case-by-case basis. Grand Central Station and Saint Patrick’s Cathedral are two examples. The size, beauty, and appearance of these buildings often make them into symbols of the city and defining features of a neighborhood’s identity.
Secondly, there are historic districts. Unlike individual landmarks of singular aesthetic value, historic districts are valuable because they form streetscapes. For instance, while individual structures in the Greenwich Village are unremarkable, together they form a unique streetscape worth preserving. A vibrant streetscrape includes structures of various ages, uses, functions, and sizes. In these districts, new development must not be much taller than and must not employ different materials from neighboring historic buildings. From the 2018 city-wide ordinance, zoning aims “to protect the character of certain designated areas of historic and architectural interest, where the scale of building development is important, by limitations on the height of buildings.” 5

.

.

Thirdly, there are, however, many non-historic and vacant parcels within historic districts. Many of the protections applied to historic buildings are extended to neighboring sites. Development on these less important sites can enhance or destroy the streetscape. For instance, most buildings in a neighborhood may retain their original appearance, but a few interspersed between were built later in a different style, or they were in some way destroyed before the area was landmarked. These structures are preserved not because of what they look like, but because of where they are located. Above are two examples.
In the case against historic preservation, contextual preservation seems the most flawed. The red tape of preservation law might disincentive needed investment in these non-contributing structures. However, fewer than 15% of all structures within historic districts are listed as non-contributing. The data is broken down below, by borough and for the city at large:

.

Borough Manhattan Brooklyn Queens Bronx Staten Island

NYC

Totals

Designated structures

(individual and districts)

32,376 28,680 25,560

17,325

 

5,344 109,285
Non-contributing structures within historic districts 6,465 3,783 2,626 3,118

1,512

 

17,504
Number of vacant parcels within historic districts 40 457 74 444 29 1,044
Percentage of buildings in historic districts that are non-contributing and/or vacant 16.731% 13.713% 9.5541% 17.054% 22.38% 14.74%
Borough totals 38,881 30,920 28,260 20,887 6,885 127,833
(all five boroughs)
Landmarks outside of any borough 761 128,594
(total)

.

This yields 128,594 6 protected structures city-wide. There are 857,271 structures total in the city. 7 which means that landmarked buildings comprise less than 14% of all structures in the city. In addition, the non-contributing buildings and vacant parcels within historic districts comprise less than 2.16% of the city’s built environment. New York City contrasts with comparable world cities like Paris and London, which are millennia older and protect a far greater percentage of their historic fabric. Below, for instance, are maps of the conservation areas in Westminster, London 8 versus Lower Manhattan and Downtown Brooklyn. 9 In other words, preservation law is limited to certain buildings and certain areas; it is too small a factor to drag down the larger city’s growth.

.

.

Study Three:

How does the preservation movement reflect economic patterns?

.

This section is divided in three parts. We first describe why civic structures are the easiest and most likely to be preserved. We then describe the economic factors why commercial structures (3.56% of all landmarks) are less likely to be preserved than residential structures (27.66% of all landmarks). Finally, we hint at possible correlations between landmarks preservation and gentrification.

.

3.1 Civic Buildings

Civic structures tend to be better preserved. New York City owns at least 14,000 properties 10 across the five boroughs. However, there are 16,920 landmarks that serve “civic” functions, including 11,726 landmarked buildings relating to public health and 571 related to armories. In fact, among all 440 types of landmarks, civic-related structures and institutions have the highest rates of landmark status and preservation.
What explains this? One explanation could be that civic sites, particularly neoclassical buildings from the Gilded Age, tend to be high-quality, well-built, and aesthetically pleasing, so as to evoke the power of government through architecture. Therefore, these buildings seem more likely to be deemed worthy of preservation.
An alternative explanation could be that civic and residential structures are easier to landmark than commercial. Elected officials are responsible for maintaining city property, and they must respond to voter and community pressure. The public can threaten to vote out officials who neglect historic, city-owned properties. Additionally, there are few reasons for developers and residents to object to preserving civic buildings.
Still yet, there are stronger factors influencing preservation. Civic structures are not subject to market pressures, and city-owned buildings do not have to help their occupants make a profit. For instance, the cost of renovating a historic public school might be more expensive. Fortunately, the city is not a profit-driven corporation. By contrast, a developer is always looking to extract the greatest profit possible from the land he owns.
Commercial structures are subject to strong market pressures favoring demolition. An old factory that has outlived its designed lifespan will be abandoned or demolished if it cannot be re-used. Converting an old factory to new uses is often cost-prohibitive, requiring environmental remediation, and lengthy approvals. If renovation cannot generate enough profit, there will be pressure to demolish. City-owned libraries and hospitals do not face this kind of pressure. This drives private developers to demolish their properties at a higher rate than public institutions, as illustrated by how few commercial structures are preserved (3.56% of all landmarks).

.

3.2: Residential vs. Commercial

With increasing land values, newer buildings are less likely to be low-density single-family homes and more likely to be high-density commercial and mixed use. However, the city seems to prefer landmarking residential over commercial structures. The table below show the building types preserved, their quantity, and the percentage of the total number of preserved buildings each building type represents. Structures are categorized by their original functions. So a building designed as a factory but later converted to residential is still listed as “industrial.”

.

Type of Building Number of Buildings of this Type Percent of Total
(rounded to .01)
Residential 35,575 27.66%
Civic 16,920 13.16%
Street Furniture 13,943 10.84%
Commercial 4,574 3.56%
Infrastructure 2,490 1.94%
Transportation 2,145 1.67%
Institutional 2,026 1.58%
Religious 1,509 1.17%
Mixed Use 1,324 1.03%
Vacant 1,178 0.92%
Military 759 0.59%
Industrial 436 0.34%
Outbuildings 11 32,391 25.19%
All other uses 14,970 11.64%
Totals 128,594 100%

.

The most important figure above is the disproportionate representation of residential and civic buildings that are landmarked. For instance, as of 2018, Manhattan has 39,172 landmarks. Of these, 35% (= 13,816) are for residential use, 9% (= 3,443) are commercial, and 1.5% (= 650) are mixed-use. Mixed use usually means commercial at ground level and residential on top. Even though more people work in Manhattan than live there, the city has preserved four times more residential than commercial structures on the island. On weekdays, 3.1 million people work in Manhattan, while only 1.6 million live here. In other words, residential buildings seem more likely to be preserved than commercial.
Our data also reveals a spatial concentration of residential buildings in historic districts. For instance, most of Manhattan’s residential landmarks are concentrated within historic districts in the Upper West Side, Upper East Side, and the skyscraper valley between Midtown and Downtown. Residential sites are more likely to be collectively landmarked as part of historic districts and streetscapes. As illustrated below, 94.93% of residential landmarks citywide fall within historic districts, and only 5.07% are outside these districts:

.

Residential All Other Types
Within historic districts 35,029 = 94.93% 61,124 = 66.66%
Individual landmarks outside historic districts 1,872 = 5.07% 30,569 = 33.34%
Total 36,901 91,693

.

What explains the disproportionate protection of residential structures? One factor could be higher income-levels in historic neighborhoods and associated protectionism (i.e. NIMBYism). The map below shows the correlation between the locations of historic districts and 2018 data on income levels and length of residence. Historic districts overlap with neighborhoods of higher incomes and longer-term residents. For instance, most residents in the Brooklyn Heights historic district have lived there for between 17.1 and 48 years, and their annual incomes range between $51,500 and $289,000. People in the rest of Brooklyn have lived at their current address for between 10.3 and 12.8 years, and their annual income is $51,500. Similar patterns play out in the historic districts of the Greenwich Village and the Upper West Side. In other words, residents in historic neighborhoods are more likely to stay-put.
Length of residency and percentage of home ownership may mirror the degree to which residents are invested in maintaining and improving their immediate built environment. The relationship between historic preservation and length of residency is too strong and too consistent across all five boroughs to be a mere accident. There may be causative factors at play between income, emotional investment in one’s community, and the willingness to fight for historic preservation. This needs to be further analyzed and confirmed with future data.

.

Launch interactive feature (opens in new tab)

Individual landmarks outside historic districts tend to be commercial structures.
There is no visible relationship between the siting of individual commercial landmarks
and the income levels of their adjacent community.

.

3.3 Gentrification?

The spatial relationship illustrated above is surprising for another reason: gentrification. Gentrification is often linked to rising living costs and the displacement of existing residents. The physical appearance of historic neighborhoods would seem to make them more desirable for gentrification. However, the average length of residency is longer in historic than in non-historic districts, even though income (and rent, too) are higher in historic districts. In other words, neighborhoods that fall within historic districts more often have high or rising incomes and longer length of residency than residents from non-historic districts. This seems contradictory because one would think that high-income areas would be more likely to displace existing residents, and therefore would be less likely to have long-term residents from the pre-gentrification era.
In contrast, neighborhoods without the benefit of historic preservation more often have more short-term residents and a high annual turnover rate. The Williamsburg neighborhood is one example with incomes over $51,500 (similar to Brooklyn Heights) but length of residency under 10.3 years. Additional research should examine if rent-stabilized apartments are more likely to be concentrated in historic districts. The legal barriers of preservation might make it more difficult for developers to push out existing residents, gut an old building, and then rebuild it to charge higher rents. Building height restrictions in these old neighborhoods also reduce the motivation to even demolish a structure to begin with because any new structure built there would not be larger and more profitable. Unfortunately, NYC Open Data has no information on the spread or geographical clustering of rent stabilized apartments in older buildings.
The possible relationship between historic preservation and gentrification needs to be confirmed through further analysis. The results of this study would indicate if historic preservation is an effective tool to stabilize neighborhoods and slow gentrification.

.

.

.

Study Four:

Keeping up to Pace?

.

When the first group of 2,312 buildings were landmarked in 1965, their average year of construction was 1882 – representing an 83-year gap between construction and landmarking. In 2018, the average construction year of newly landmarked structures is 1908, representing a 110-year gap. Thus, in the 53 year life of the landmarks movement from1965 to 2018, the average age of a building when landmarked has increased by 37 years.
The more recent inclusion of modernist skyscrapers, like the Lever House (completed 1952 and landmarked 1982) and Seagram Building (completed 1958 and landmarked 1989), may give the impression that the criteria for what qualifies as important and worth preserving has expanded. Our data does not support this conclusion, because while recent years have seen some newer buildings granted landmark status, the rate of designation has not kept up with the rate of construction and, in fact, has fallen behind.
The graph below illustrates – for a sample size of 5,451 structures – the date a structure was landmarked on the horizontal axis measured against its construction date on the vertical axis. Structures are plotted on this graph by color. Individual dots represent individual sites. The black trend line indicates that between 1965 and 2018, the average age of new landmarks has only slightly increased. The buildings the city is protecting today are only slightly newer than the kinds of buildings being protected in the 1960s.

.

5,451 records above

.

Is the scope of historic preservation limited to the ninteenth-century? Since 1965, thousands more buildings have become eligible for landmark status, but they are not often protected. Is the city no longer building the kinds of structures deemed worthy of preservation?
To qualify for landmark status in NYC, a building must be older than 30 years (or older than 50 if added to the National Register of Historic Places). From a publication by the The Trust for Architectural Easements: “LPC property must be at least 30 years old – no exceptions – whereas a National Register property must be at least 50 years old, unless it is found to be of exceptional significance, in which case there is no age limit at all.”  12
When the Landmarks Preservation Commission was formed in 1965, none of the buildings from 1935 to 1965 qualified for protection. Today, as of 2018, any building built before 1988 can qualify. However, less than 5% of all listed structures date from the 53 years from 1935 to 1988. This was a significant and long time in this metropolis’ history, but the architectural record from this time is not well landmarked.
The graph below illustrates – for a sample size of 5,451 structures – the distribution of landmarks by year built. On the horizontal axis are the years built from the 1600s to the present-day. On the vertical axis are the estimated number of landmarks built in each year, and which are now protected. Most buildings fall within the ninety year span from 1850 to 1940, peaking in 1895. Few landmarks fall outside this time period.
The rise and falls on this graph may also correspond to the roughly twenty year cycles of boom and bust recessions, along with corresponding halts in new construction. The shortage of pre-1850 sites is explained by how the city was smaller before 1850, and therefore had fewer landmarks to begin with. However, the shortage post-1940 landmarks may hint at a larger historical oversight on the part of historians and city government.

.

93,691 records above

.

The 1965 landmarks law was part of the city’s response to the demolition of old New York Penn Station in 1963. Countless significant buildings had been lost to urban renewal in the name of progress. Activists wanted to prevent continued destruction. By the 1960s, urban renewal was winding down. New York was entering the prolonged recession of the 1970s and 80s, during which urban renewal and new construction ground to a halt. In this light, landmarks law originated as a post-facto response to demolition that had been going on for decades.
Must landmarking occur after destruction of newer landmarks has already begun? There are doubtless hundreds of post-war significant buildings that have not yet been identified or deemed worthy of preservation. The question is not: Should we list these buildings? Rather, the question should be: Why are we not listing these buildings before they are threatened? The Museum of Modern Art’s 2014 decision to demolish the American Folk Art Museum is one example of a recent building that could, or should have, been landmarked so as to prevent demolition. 13

.

Conclusion:

The Future of Historic Preservation

.

Data can only reveal so much. These statistics do not speak of specific historic sites or the unique identity of each. This data, however, reveals big picture trends, biases, and possible problems with historic preservation. These trends are invisible from street level or at individual sites; they are only visible through the lens of data. From this data-driven analysis, we draw four main conclusions:
Firstly, preservation law is subject to political pressure. The geography of historic preservation seems to preference some neighborhoods (usually higher incomes ones) over other neighborhoods with lower incomes. Preserving and restoring old buildings takes effort and money.
Secondly, many developers accuse historic preservation of slowing new construction and economic growth. Yet, landmarked buildings comprise only 14% of the city’s buildings, while non-contributing structures within historic districts comprise only 2% of all buildings. There is ample room for new development outside historic districts; development pressures on landmarked areas can be directed elsewhere.
Thirdly, residential properties are preserved in disproportionately greater numbers than commercial and industrial structures. The community and economic pressures to redevelop are different for different types of buildings. Most residential landmarks also fall within historic districts, and are therefore parts of the urban streetscape. Residents often use preservation law to protect their streetscape and the homes they own from new development that would weaken property values. Neighborhoods of lower-density old buildings, like the West Village, retain their popularity, charm, and high property values thanks to strong legal barriers against new development. Absent these protections and legal guarantees, property values could depreciate.
Linked to this third observation, the market pressures to demolish civic structures are weaker than the market pressures to demolish commercial and residential. As a result, a higher percentage of city-owned or institutional buildings are preserved, and a lower percentage of commercial and industrial.
Fourthly , historic preservationists prefer to protect pre-WWII buildings, even though numerous post-war examples qualify. As a result, there are a high number of prewar buildings with landmark status, and comparably fewer postwar landmarks. Similarly, the rate at which landmarks are designated has not kept up with the pace of new construction.
The economic success of New York on a global scale and its continuing construction boom caused the demolition of many non-residential commercial landmarks that would have otherwise qualified for landmark status had there been fewer pressures for economic development. In the words of leading NYC historian, Kenneth Jackson:
History is for losers. By that I mean, cities which have chosen to preserve all their historical monuments and locations usually do so because no one else wants the land to develop. Modern progress has passed them by. New York’s history doesn’t litter the streets visually, it can be hard to find sometimes, but that is because the city is an economic winner on a global scale. 14
New York is indeed a winner “on a global scale.” While Wall Street symbolizes America’s economic power, the United Nations symbolizes America’s political power. The city’s over three million foreign born shape the city’s identity as an interconnected and diverse metropolis.
Nonetheless, progress has an aesthetic cost, as reflected in the countless lost landmarks and in Midtown’s dark and monotone skyscraper canyons. Fifth Avenue’s Gilded Age mansions and old Penn Station are gone; so, too, are the picturesque skylines and distinctive ethnic neighborhoods of Berenice Abbott’s 1930s photographs. New York is different today. While streets and subways grow more crowded, climate, flooding, and tropical storms threaten the city’s fragile ecology and outdated infrastructure.
It is too early to judge whether the city is architecturally poorer or richer for progress. Although historians discourage speculation about the past or alternative histories, how would the political or cultural landscape of New York be like today without landmarks law? This, however, is a question data cannot answer.

.

Links to Resources

The original datasets can be viewed or downloaded below:

.

Further Reading

Anthony Wood. Preserving New York: Winning the Right to Protect a City’s Landmarks. New York. Routledge. 2008.

.

Endnotes

  1. “Individual Landmarks,” NYC Open Data, https://data.cityofnewyork.us/Housing-Development/Individual-Landmarks/ch5p-r223 (retrieved 5 November 2018).
  2. “LPC Individual Landmark and Historic District Building Database” NYC Open Data, https://data.cityofnewyork.us/Housing-Development/LPC-Individual-Landmark-and-Historic-District-Buil/7mgd-s57w (retrieved 5 November 2018).
  3. NYC’s 2017 population is an estimated 8.623 million.
  4. More on this topic: Rachel Mollie Levy, “Contextual Zoning as a Preservation Planning Tool in New York City,” (Master’s diss., Columbia University: Graduate School of Architecture, Planning, & Preservation, 2015) https://academiccommons.columbia.edu/doi/10.7916/D8HD7TVM (retrieved 5 November 2018).
  5. “General Purposes of Residence Districts,” in The Zoning Resolution: Web Version, (published by NYC Zoning Department, 2018), pp.252-53. https://www1.nyc.gov/assets/planning/download/pdf/zoning/zoning-text/allarticles.pdf (retrieved 5 November 2018).
  6. The total for all five boroughs is 127,833. Including landmarks not registered in any borough, like Ellis Island, the total is 128,954.
  7. NYC Planning Department, “Spatial Data Properties and Metadata,” from MapPLUTO, (published to the web, 2018), pp.5 https://www1.nyc.gov/assets/planning/download/pdf/data-maps/open-data/meta_mappluto.pdf?v=18v1 (retrieved 5 November 2018).
  8. “Conservation Areas,” City of Westminster, https://www.westminster.gov.uk/conservation-areas (retrieved 5 November 2018).
  9. Published by NYC Zoning Department, “NYC_Historic_Districts_2016,” ArcGIS 9geographic information system), https://data.cityofnewyork.us/Housing-Development/Historic-Districts/xbvj-gfnw (retrieved 5 November 2018).
  10. “New York City owns or leases 14,000 properties around the five boroughs—a public asset with the cumulative surface area of Brooklyn.” From: “Public Assets: Mapping the Sixth Borough of New York,” The Municipal Art Society of New York, https://www.mas.org/initiatives/public-assets/ (retrieved 5 November 2018).
  11. “Outbuildings” include garages, stables, street furniture, and accessory structures. This category skews the data table. Since many accessory structures were turned into residential structures, the actual percentage of current residential dwellings is higher than 27.66%.
  12. Anthony W. Robins, “Differences between Landmarks Commission Designations and National Register Listing,” in Similarities and Differences between Landmarks Preservation Commission Regulation and Donation of a Preservation Easements, (Prepared for The Trust for Architectural Easements, 2009), pp.10, http://architecturaltrust.org/~architec/wp-content/uploads/2013/06/1a-2009-0512-Robins-Report.pdf (retrieved 5 November 2018).
  13. Michael Kimmelman, “The Museum With a Bulldozer’s Heart,” The New York Times, January 14, 2014, https://www.nytimes.com/2014/01/14/arts/design/momas-plan-to-demolish-folk-art-museum-lacks-vision.html (retrieved 5 November 2018).
  14. “Quotes from Kenneth Jackson,” CULPA, http://culpa.info/quotes?professor_id=97 (retrieved 5 November 2018).

The Geography of Art History

According to the Metropolitan Museum of Art

.

Related: Data analysis and visualization of 120,000 works in the Museum of Modern Art

.

.

In this film, each colored dot indicates one location represented by art in the Met’s online database. Dot location indicates artwork provenance. Dot size indicates the number of objects from this place. The time each dot appears corresponds to the year this work was created. This data is assumed to be an accurate sample size.

.

Over the past few years, the Metropolitan Museum has catalogued over 25% of its holdings online. This represents ~590,000 objects, covering over 5,000 years of human history from 17 curatorial departments. The diversity of objects in a museum’s collection (and the amount of contextual information known about these objects) may reflect the kinds of narratives a museum can curate about artistic and global history. This visualization charts the provenance and year of production of every single object that is catalogued on the Metropolitan Museum website, whenever this information is known.
The geography of art history is uneven. Certain regions, particularly cities, are home to diverse and famous artistic output. Thomas Friedman similarly describes globalization as being spiky and concentrated in big cities. Other regions are comparatively less productive and less studied. Either this reflects museum curators’ historic bias against Africa, Latin America, and the “Global South” in favor of Europe. Or this might reflect a more fundamental historical reality: If geography guides artistic production and privileges regions with good geography, like areas surrounding the Mediterranean, then landlocked and inaccessible regions with poor geography will have less artistic output.

.

.

Art objects from ancient cultures like China, Egypt, and Sumeria frequently have known provenance but unknown year of production. Unfortunately, they are therefore excluded from this visualization. There are many objects in the collections with known provenance but unknown production date. Figure one illustrates objects with known provenance and known year. Figure two shows objects with known provenance only.

.

.

The original data was downloaded here from the Met Museum’s website.
This visualization and interactive map are free to view and download here.

The University of Oxford in a Box

 

.

Music: Minuet in G Minor by Bach

.

This paper model of Oxford folds out of a leather suitcase. This is a souvenir of my Oxford experience.
Half the model features university’s main buildings: The Radcliffe Camera, Bodleian Library, Sheldonian Theatre, University Church, and the Clarendon Building. The other half is the campus of Saint Catherine’s College in Oxford.
Attaining precision in a small model is difficult and took about two weeks to build. It is made from paper cutouts, which were folded into various shapes. Below are two of the cutouts before assembly. Dimensions: 7 by 14 inches and 1.5 inches deep.
Below is the image of one of these cutouts before assembly, and the ground plan of the campus before the paper buildings were mounted on cardboard.

.

 

.

.