Batty Computational Urban Science (2022) 2:8 Computational Urban

https://doi.org/10.1007/s43762-022-00036-z
Science

ORIGINAL PAPER Open Access

The Linear City: illustrating the logic

of spatial equilibrium
Michael Batty*

Abstract
Linear cities where activity is spread out along a transportation line, aim to offer the highest levels of accessibility to
their adjacent populations as well as to the countryside. These city forms are popular amongst architects and planners
in envisioning ideal cities but they are difficult to implement as they involve strict controls on development which
often ignore human behaviour associated with where we locate and how we move. We briefly explore the history of
these ideas, noting the latest proposal to build a 170 km city called Neom in north west Saudi Arabia, a plan that has
attracted considerable criticism for its apparent ignorance of how actual cities grow and evolve. We use a standard
model of human mobility based on gravitational principles to define a set of equilibrium conditions that illustrate
how a theoretical city on a line would, without any controls, successively adapt to such a new equilibrium. First, we
represent the city on a line, showing how its population moves to an equilibrium along the line, and then we gen-
eralise this to a bigger two-dimensional space where the original line cutting across the grid evolves as populations
maximise their accessibility over the entire space. In this two-dimensional world, we simulate different forms that
reflect a balance of centralising versus decentralising forces, showing the power of such equilibria in destroying any
idealised form. This approach informs our thinking about how far idealised future cities can depart from formal plans
of the kind that the linear city imposes.
Keywords: Linear cities, Neom, Gravitational interaction, Accessibility on a line, Spatial equilibrium, Developments on
a grid

Cities that capture the ideas of visionary architects and Cities tend to grow around some centre in concentric
planners often bear little resemblance to the way most zones of land use sorted according to their ability to pay
cities grow and evolve. Lucio Costa generated his plan rent, linked to this core using well-defined radial routes
for the new capital of Brasilia by drawing a ‘cross’ on the converging on the centre. Since Brasilia was founded, the
legendary ‘back of an envelope’. He proceeded to elabo- population and its various activities attracted to the city
rate it into the shape of an aeroplane (or bird) which then have located more or less anywhere they can around the
became the master plan around which everything else ‘cross’, generally ignoring the dominant form of the mas-
revolved as the city was rapidly constructed from the late terplan which consists of the two wings of the plane or
1950s on. If you examine this sketch, you are immedi- bird that Costa used to define its basic form. We show the
ately struck by how unrealistic it is and on thinking about plan and its origins in Fig. 1.
how people actually behave in cities, it looks as though In the 70 years that has elapsed since the plan was
this shape is completely incompatible with what we know begun, the city has evolved to an entirely new equilib-
about the form and function of most contemporary cities. rium more like a contemporary city of polycentric form.
Costa’s masterplan is still enshrined in the two wings but
most of the city resembles a much more naturally grow-
*Correspondence: [email protected] ing organic form (Banerji, 2012). If this lesson in archi-
Centre for Advanced Spatial Analysis, University College London, 90 tectural determinism means anything at all, it suggests
Tottenham Court Road, London W1T 4TJ, UK

© The Author(s) 2022. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which
permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the
original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or
other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this
licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.
Batty Computational Urban Science (2022) 2:8 Page 2 of 17

Fig. 1 Brasilia. A) Lucio Costa’s Original Sketches, B) his first Master Plan, and C) Configured Geometrically as a Plane or Bird and D) A Satellite View
of the Original Wing Structure, Today, from Google Earth

that cities have a natural equilibrium which will always will have no cars but a hyperloop system that is able to
assert itself once the population is able to make its own transport anyone to any place on the line in no more than
locational decisions over where they might live and work 20 min. In essence, the city is composed of ‘three lay-
and how they might interact with one another. You might ers’ or ‘tubes’ – three lines: a walking layer, below that a
think that experiences like this convince us that ideal high speed transportation layer that moves people using
cities simply fashioned on regular geometric forms that the hyper-loop technology, and a third, deeper layer that
bear no relationship to the way the city might function contains services and utilities. The city which is called
would have been abandoned long ago. But far from it. In Neom (‘New Future’) or more colloquially The Line, will
2020, the Crown Prince of Saudi Arabia proposed a new be carbon neutral, economically vibrant with employ-
city for a million people who are to be spread out along a ment based on high tech, and pollution-free. So little has
dead straight line 170 km in length.1 It will be a city where been worked out however, that its plan so far is no more
everything one needs will be within five minutes’ walk, than a line across the region of Tabuk in the northwest
it will be no more than half a kilometre in width, and it corner of the country (see https://​www.​neom.​com/​en-​
us/​whati​sthel​ine). Needless to say, it has attracted severe
criticism as an indulgence by the rich and powerful.
1
See video of the plan for Neom https://​www.​youtu​be.​com/​watch?v=​41sgR​
P0G6y4
Batty Computational Urban Science (2022) 2:8 Page 3 of 17

If Brasilia was unable to adhere to the form imposed by not already populated are considered for growth. These
its masterplan, then how do you suppose development in alternatives lead to city forms that are much more likely
Neom can be controlled to rigidly keep to the Line; only to resemble our contemporary cities in terms of shape
by severe restrictions that would limit the most basic and polarity, and essentially demonstrate the futility of
freedoms. In contemporary cities, people move accord- building cities whose form ignores the functions that
ing to the time, cost and distance incurred by visiting determine how we locate and move in the real city.
activities removed from places where they live and work.
Tobler’s (1970) law states “… everything is related to eve- 1 The history of the Linear City
rything else, but near things are more related than distant Ideal cities have been proposed by philosophers and
things …” , and it operates everywhere. In Neom, unless architects since antiquity and they all share the common
you are forbidden from actual travelling to meet anyone characteristic that they are composed of relatively simple
or from interacting with anything at different distances geometries, most reflecting the straight line, the grid, the
away from you which would be impossible, some places circle, and just occasionally the flowing line. Most have
near the centre of the Line will be much more accessible been new cities which begin small but as they grow and
than those at each end. People and activities will begin develop, lose their ideal form as a multitude of locational
to adjust with respect to their ability to pay for different decisions made by the population crowd in on the origi-
levels of accessibility and the Line will become differenti- nal plan. Greek and Roman castra are perhaps some of
ated with respect to the value of different locations. The the clearest examples where the residuals of these grid
ultimate form of the city will thus be no longer uniform iron plans used to organise the structure, are still evident
and linear. at a very small scale in cities that have grown dramati-
We will begin by sketching what we know about ideal cally since they were first established as encampments.
cities based on linear forms, and how these forms have The extreme form of this idea is the linear city where like
been almost impossible to maintain as they collapse Neom, population is spread out along a line which invari-
towards a relatively stable equilibrium. We will then ably provides a transportation artery enabling interaction
propose a simple geometric representation of an ideal with both near and far neighbours who live on either side
linear town and then suggest how population units that of the line at different distances away from each other
occupy the form interact with one another using a vari- but adjacent to the country and agriculture which some-
ant of Tobler’s Law. This suggests a process whereby the times provides their livelihoods. The classic example was
population units move towards sites of greatest accessi- proposed in 1882 by Arturo Soria y Mata who suggested
bility which is akin to introducing distance between those such a linear form configured as a horseshoe shape
parts of the city that define adjacent units spanning the encircling much of Madrid, organised around a tram-
line. This enables us to state a formal equilibrium that way not more than 500 m wide so that it could grow in a
shows us how a stable form emerges which we demon- straightforward manner at either end (Boileau, 1959). In
strate first for the line, maintaining its integrity but allow- the event, only a 5 km segment was built with the devel-
ing populations to move to any of the cells defining the opment stopped in the 1930s by the Spanish Civil war.
units which people occupy along the line. This process Today La Ciudad Lineal as it is called is absorbed into the
retains the linearity of the city but changes the numbers sprawl of greater Madrid, in much the same way the plan
of population in each of the locations. The outcome is to for Brasilia has become a residue, perhaps a victim from
an extent obvious. The uniform distribution of popula- an earlier era as the growth of the city has surrounded
tion is transformed to one where population (and its den- it. In fact, Soria y Mata considered such linear develop-
sity for each unit has the same area) is proportional to the ment the obvious and ‘optimal’ way to grow a city, echo-
accessibility of each location. ing much later ideas about cities as fractals when he said:
We then relax this assumption that the only permissi- “The growth of linear cities is simple because the line
ble locations lie on the line. We assume the line defines itself can go on to infinity and at any point a new commu-
an axis across a space which is a grid of locations. The nity can shoot off like the branch of a tree, the tributaries
initial city is the uniform population along the line and of a river, the veins of the body“, quoted in Furundzic and
we then compute accessibilities across the entire grid. Furundzic (2013).
Locations which are not part of the line have accessi- The first 30 years of the last century saw many propos-
bilities and we then begin to populate the whole space als to develop linear settlements, also referred to as strip
with respect to adding population in proportion to these cities, ribbon development, and corridors. Tony Garnier’s
accessibilities. We then modify this to only consider new Cité Industrielle reflected his ideas about the optimal city
populations with respect to the highest accessibilities and which suggested strong linearity of form (Wiebenson,
then we examine the situation where locations that are 1960). The emergent Soviet Empire also had its fair share
Batty Computational Urban Science (2022) 2:8 Page 4 of 17

Fig. 2 A Sample of Linear Cities at Different Scales and Times

of adherents to the linear city ideal which it was argued architecture and then through his Athens Charter. This
provided a sufficiently flexible form to embrace contin- had an enormous influence on the design of public hous-
ued relentless expansion of manufacturing production ing after the second world war but he had an impact on
lines as in Nikolay Milyutin’s (1930) plan for the social- one of the most radical linear city ideas ever proposed
ist city. These ideas were outlined in his book Sotsgorod by the Modern Architectural ReSearch group (MARS).
and in many proposals for Soviet cities such as in Ernst Corbusier’s thinking also permeated the Bauhaus and
May’s plan for Magnitogorsk. In particular, Milyutin’s it was Ludwig Hilberseimer who proposed linear-like
plan for Stalingrad used the natural features of the Volga residential structures which influenced the architects in
to impress his ideas about linear growth in the form of a MARS. The group which evolved itself through various
‘strip city’ using a strict parallelism of zones of different ideologies about the form of future cities, produced a
land use (Antyufeev and Antyufeeva, 2019; Collins, 1959; plan for London whose form was to be based on 8 radial
Doxiadis, 1967). Many of these ideas were taken up by strips, north and south of the Thames making 16 strips
groups of architects who saw the city more as a work of in all, a little like taking Soria y Mata’s La Cuidad Lineal
art than a living system, more as a machine that could be and splaying it north and south of the River, replicating
designed using simple principles of interaction and loca- it 16 times. The plan which was ultimately published in
tion than an organism whose own forces held its dynamic 1942 contrasted markedly with how the London actu-
structure together. ally functioned and in hindsight, it almost represented
In the 1920s, the architect, Le Corbusier, one of the an ongoing thought experiment in fictional geometries
pioneers of the modern movement, also began to spec- for the future city. Gold (1995) provides a fascinating
ulate on the shape of idealised cities beginning with summary of the social dynamic in preparing the the plan
strongly radial and circumferential forms but rapidly over a period of some 10 years. We show a sample of
evolving his ideas into a much stricter geometry – the these linear forms in Fig. 2.
linear. With Soria y Mata’s La Cuidad Lineal as inspira- The properties of these ideal cities were not well-
tion, he devised what he called the Ville Radieuse into a worked out by their proponents. Apart from the
city form that mapped various geometric lines and the general notion that a linear structure directs trans-
location of key urban activities into a shape that reflected portation that can be easily accessed if development
the disposition of functions within the human body is kept to within walking distance of the line or spine,
(Corbusier, 1935). He aggressively promoted the idea and that the countryside is equally accessible on either
through several of the CIAM conferences on modern side of the line, there is rarely any discussion of overall
Batty Computational Urban Science (2022) 2:8 Page 5 of 17

accessibility. However several examples do suggest that This brings us back to Neom. In Fig. 2, we illustrate
such lines can join more conventionally configured the only plan that appears to have been published for
radially concentric clusters together, thus increasing the city although it is not possible to picture any detail
overall accessibility. In fact, March (1967a, b) argues of development over the 170 km span. Neom depends on
that if one assumes a central place hierarchy where the two key principles: the latest technology to enable the
hubs simply represent points where linear strands of city to remain connected across its length; and access to
development come together, then rather than focus- open space from within the modules or the clusters that
sing all the development on the hubs, it can be spread make up what essentially might be like a series of beads
out over the lines joining the hubs. In this way, the on a necklace. These two principles are consistent with
proposed form increases the accessibility of each unit the move to net zero carbon emissions and a healthy life
of development with respect to any location within the style through people interacting using active travel to
structure but this argument also depends on the density local amenities and facilities. The unsustainability, how-
of development and on the width of the linear strands. ever, of Neom is that no other principles relating to loca-
Nevertheless, as March (1967b) points out, this style of tion, interaction and density are invoked. The hype that
ribbon development does have advantages. He says: has accompanied its launch (even advertisements on UK
commercial TV) suggests many aspects of the future
“I could argue that the forces behind the most
city that are mere aspirations. The brochure says: “It’s an
spontaneous and energetic form of urban growth
attempt to do something that’s never been done before
to have occurred this century in this country,
and it’s coming at a time when the world needs fresh
namely ribbon development, are seen to be not
thinking and new solutions … part of building a new
only rational, but economical in respect of roads
model for sustainable living, working and prospering”.
and services. And that, paradoxically, if these
Yet the denunciations have come thick and fast. Gor-
forces had been directed to create planned rib-
don (2021) says:
bons, the feeling of builtupness in the country
as a whole could have been less, not more, than “If techno-futurism has perfected anything, it is the
today’s emphasis on urban envelopes will produce” art of unwittingly re-inventing old ideas, inflating
(March, 1967b, page 335). them to a scale so epic that it accentuates all of the
idea’s flaws, and presenting it in a slick hype video as
There are other linear structures that have been used to
the Only Way Forward.”
describe contemporary as well as ideal cities which focus
not on cities built along a simple line but on many lines. Worth (2021) writing in the New York Times says:
Space syntax has been developed to enable cities to be
“Now the Saudis have announced a fantasy that
represented as constellations of lines and Hillier (1996)
makes all their previous efforts look tame …” calling
has proposed several idealised configurations which are
the Line “… ‘a civilizational revolution’ to be inhab-
used to define streets. There are many discussions of
ited by one million people ‘from all over the world’.
streets and patterns by various urbanists, see for example
Why anyone would want to move there, and why a
the work of Marshall (2005) amongst others, while more
city should be shaped like a strand of capellini, is
formally the whole body of network science which has
anyone’s guess.”
developed rapidly over the last 20 years, is relevant to the
broader discussion. At one extreme, there are examples Let us give the last word in this rapid historical per-
of cities without streets at all, largely in the oldest urban spective on the linear city to Jane Jacobs (1961). Her the-
settlements such as those in Asia Minor and Mesopota- sis has always been that cities are messy affairs, highly
mia. At Catalhoyük in Turkey, there is what archaeolo- diverse, full of essential tensions and contradictions that
gists consider a map of the settlement that has no streets give them their true vitality. In speaking of idealised
and where the current interpretation is of a place where forms, the brain child of those architect-planners who
the inhabitants worked out of doors on an elongated believe it is their role to pronounce, she says:
deck of sorts into which were punctuated the various
“When city designers try to find a design device that
dwellings (Hodder, 2005). The plan of the village shown
will express, in a clear and easy fashion, the ‘skele-
in Fig. 2 is probably the world’s oldest map despite some
ton’ of city structure (expressways and promenades
controversy (Clarke, 2013). Catalhoyük is over 9000 years
are current favourites for this purpose) they are on
old and its origins came well before the world invented
fundamentally the wrong track. A city’s very struc-
wheeled technologies, thus impressing on us the power
ture consists of a mixture of uses… We get close to its
of technology in suggesting and guiding the development
structural secrets when we deal with the conditions
of various urban forms.
Batty Computational Urban Science (2022) 2:8 Page 6 of 17

Fig. 3 Formal Representation of the Linear City

that generate diversity” (Jacobs, 1961, page 376). further, we need to explore how its cells interact with one
another.
Diversity of course is not what cities on a line are capa-
The ability of any two cells to interact is also a key fea-
ble of generating, at least as long as the line is strictly
ture of this model but in all the examples of linear cities
maintained. We will now demonstrate this formally.
that we introduced earlier, such interactions are miss-
ing, assumed to be zero or simply not stated. Indeed if
there are any assumptions at all, it is that such interac-
2 A spatial representation of the Linear City
tions are constant with relative distance between the cells
We begin with an assumption that a linear city is a set
that make up such linear structures not playing any part.
of regular units or cells strung out horizontally and uni-
Here however distance is a key feature of our representa-
formly in sequence, like beads on a string. The devel-
tion and we thus proceed as follows. Each cell is defined
opment of such a structure takes place with the city
by its locational index xz and the Euclidean distance dxn xm
expanding in both directions from a central seed site that
between any two locations xn and xm is stated as follows:
we define by the coordinate index xo. The progression
that we assume which is the growth path of the city, adds
two cells at time t, each cell being on opposite sides of

if xn ∈ i and xm ∈ j, then dxn xm = i + j 
the central seed site. These cells are indexed as xi in one



direction and xj in the other where these indices are not if xi > xi′ then dxn xm = i − i′ (2)
only the locations of the cells but are associated with the



if xj > xj′ then dxn xm = j − j ′

time when the cells are first developed. In short, the cells
develop in sequence as xi, i = 1, 2, 3, …. . T and xj, j = 1, 2,
In defining distance between any two cells, we can
3, …. . T where i and j are also t, and T is the total number
assume that this relates to the interaction between them
of time periods defining the growth process. We illustrate
which we define as an inverse function
this sequence in Fig. 3.
We will generate a succession of two cells each con-
taining one unit of population defining two households Txn xm ∝ f −1 dxn xm (3)
 
which we call Pi(t) and Pj(t). The growth process is
Equation (3) lets us compute
 a −1 simple accessibility or
centrality for each cell as that shows
 
m f d xn xm
P(t) = Pi (t) + Pj (t) + 1 = t + 1, i, j, t = 0, 1, 2, 3, . . . , T ,
how these cells might be differentiated. We will develop
(1)
a number of these later in the paper but here we will
where λ is the growth rate and unit 1 in eq. (1) is the state a particularly simple measure based on the inverse
seed household at xo. In Fig. 3, the first variable Pi(t) square law that defines the accessibility of a cell xn. This
determines the cell xi on the left-hand side of the source is defined as
xo and the second variable Pj(t) the cell xj on the right-
hand side of xo. The linear growth rate λ = 2 is simply 
based on adding two cells at each time period, the one An = 1/ dx2n xm (4)
on the left hand side, the other on the right hand side. m
From the start of the process at t = 0, the sequence of
and it is easy to see from these accessibilities which we
household populations generated is 1, 3, 5, 7, 9, 11, ….
scale and illustrate in Fig. 4 that the central cell xo has the
This is the simplest growth process consistent with a
highest accessibility Ao.
city linear in space and time; we will adapt this to a more
We can also assume other growth processes where
complex nonlinear process below but to take this model
the number of cells generated at each time period more
Batty Computational Urban Science (2022) 2:8 Page 7 of 17

Fig. 4 Accessibility of the Original Linear City Based on An ∝ 1/ d2xn xm


m

than doubles if we increase the rate λ, noting that to the bottom up, in fact, as most cities do which we will
generate whole households the rate should be integral, show here.
that is λ = 1, 2, 3, …… If the rate is real, then it is neces- We now need to dimension the linear sequence of
sary to introduce rules that ensure the household and cells so that it accords to what is as close as possible to
cell are discrete, and if a different non-linear growth the proposed shape. In the case of Neom, the proposal is
process is specified such as P(t) = tλ + 1, then we need for the city to be built on a more or less straight-line of
to define rules that convert the cells into an exponen- 170 km in length, eventually with a population of 1 mil-
tially growing sequence of cells located on each side of lion. If each individual were to be arrayed along this line,
the central seed site x0. In this paper however, we will they would be compressed shoulder to shoulder at an
focus on the simplest model which is linear in terms of interval of about 20 cm between each which is extremely
growth, both spatially and temporally. congested. If we assume that a household were composed
However any serious model of urban structure must of an average of 4 persons, then the households would
also adopt the principle that populations are intrinsi- be spaced 2/3rds of a metre apart. In fact the proposal is
cally mobile with respect to a) daily patterns of move- that households be aggregated into neighbourhood mod-
ment such as journeys to work, shop, entertainment ules which are composed of four clusters, and although
and so on as well as to b) movements over longer peri- the size of these clusters and modules is not specified in
ods of time as populations adjust to the changing urban the Neom proposal, we will assume that a cluster con-
landscape that provides them with various opportuni- sists of 150 households with each module composed of
ties for housing and jobs. These involve more perma- 600 households or 2400 persons. If we array the modules,
nent changes in location. We have already noted that then there would be about 420 along the 170 km line and
we can define a volume of spatial interaction as Txn xm they would then be spaced at about 400 m between each.
which depends on distance or travel cost between dif- Immediately you can see that populating the line in such
ferent locations. In fact, in the simplest linear cities a way leads to unacceptably high densities if the module
from La Ciudad Lineal onwards and particularly for the is compact and constrained within the space available. It
recent proposal for Neom structured along The Line, does not appear that the geometry of the city has been
interaction between the cells or locations is assumed accounted for with respect to these constraints.
to be based on a uniform friction of distance; in other To proceed with a structure like Neom, we will assume
words, populations are assumed to interact regardless that the line is divided into 400 modules each with an
of how distant they are from one another, even though internal circulation pattern based on walking. To ensure
the hyperloop technology in Neom would still embody we are dealing with round numbers, we assume the line
a friction of distance. The implication is that the inter- is now 160 kms and thus the 400 modules are spaced at
action potential is the same from any cell to any other, some 400 m between each, nearly half a kilometre, and if
that is, that the notional distance dxn xm is constant or it is assumed the unit cell is square, then the area of the
even zero. Thus Tobler’s (1970) Law does not apply. cell is 0.16 ­km2. If each unit has 2400 persons, the den-
This almost smacks of teleportation, but in essence, the sity is about 15,000 persons per square kilometre. Lon-
notion that people will travel in the way assumed is an don has 5000 while Shanghai has about 13,000. In fact
extreme fiction, and if any such city were to be built, much depends on the area taken and in this calculation
it would soon begin to unravel and collapse to the sort we have not assumed any other land take within the city.
of urban forms that characterise cities that evolve from In practice even in a place like Neom, the densities will
Batty Computational Urban Science (2022) 2:8 Page 8 of 17

be much lower than this. However we will make use of population Ps(1) = 1, ∀ s, at the first time t = 1 is uniform
this template of 400 cell units for 2400 persons in each across all cells. In fact in real cities, accessibility defined
giving some 960,000 persons spread over a line which in these terms is usually highly correlated with popula-
is now 160 kms in length. In this sense, the structure is tion. The implication is that if the constraints on the loca-
linear but the beads along the string are now integrated tion of population were to be dropped in linear cities
neighbourhoods that we treat as indivisible cellular units. like Neom, the population would quickly adjust to a new
equilibrium.
3 Defining a spatial equilibrium This is the core relation in our analysis of cities in equi-
We will assume that the 400 modules that compose the librium and we define it from the following equation
linear city are represented by 200 modules defined by
their locations xn to the left of the seed site xo and 200  
to the right xm. We define the interactions between these

−α −α
Ps (t + 1) = P
r
Ps (t)drs /
s
Ps (t)drs (6)
locations as Txn xm and the distances between them as
dxn xm but we need to refer to these as any pair of loca- Equation (6) illustrates that the equilibrium is based
tions over both left and right series of locations which on a simultaneity between population in the different
run from x200…x1, x0, x1… x200; thus we can define any cells and their accessibility. If an equilibrium exists, then
two locations across this range as r and s. We now state Ps(t + 1) = Ps(t) and this can be derived by starting the
the generic flow Trs(t) from eq. (3) as iterative sequence from Ps(1) = 1, ∀ s. From the data we
have specified, we have explored this for the 400 module
−α linear city2 first using an inverse distance law with differ-
Trs (t) ∝ Ps (t)drs , (5) ent values of α, and then computing the distribution of
where t is time (or iteration), Ps(t) is the population in population from eq. (6). In fact it appears that the city is
cell s at time t, and α is the friction of distance parame- almost in equilibrium as soon as the movements set off
ter. In most linear cities including Neom, the assumption to meet the equilibrium condition in eq. (6) are initiated
is that drs
−α is a constant or in fact is zero, thus implying whereas the shape of this equilibrium is most sensitive to
there is no physical variation between cells r and s, and the parameters of the inverse power function. In Fig. 5,
in the case where there might be physical movement, the we show the equilibrium computed by iteration on eq. (6)
population is initially Ps(1) = 1, ∀ s. The total volume of for the functions drs−0.5, d −1, and d −2 and we also contrast
rs rs
travel where it physically takes place at any time is con- this with a negative exponential function exp(−0.1drs)
strained to the total population, that is ∑r∑sTrs(t) = P. which converges much faster to an equilibrium than the
The key assumption in Neom is that there are no indi- power functions.
vidual movements other than walking within modules or It is clear from the progression of moves, that the most
between them but if any distance is to be traversed, it is accessible point – the seed site – attracts more and more
to be accomplished by hyperloop technology. This as one population and although we have not explored this for-
might imagine is unspecified for the physical constraints mally, we are guessing to the ultimate outcome being a
on where one might access the loop are not worked singularity that is obvious from the geometry. This how-
out. In fact, if the only way to travel any distance other ever is an almost fictional interpretation of the equilib-
than by walking is using this technology, then the access rium involving the process of starting with a uniform
points to the hyperloop need to be clustered at the loca- distribution of population mandated by the designers
tions of the neighbourhood modules and there are far too of the city. The process begins where people move in
many of these (400) for this to be possible. Most of the response to the variations in accessibility which pro-
time travelling would be in accessing the trains and in duce a new distribution of population, i.e. Ps(2) ≠ Ps(1),
any case, with current technologies, it would not be pos- Ps(1) = 1, ∀ s and it continues until an equilibrium of
sible to keep to appropriate headways (Hanson, 2020). In some sort emerges. This process of continued relocation
short, the proposed basis for long distance travel in such is likely to occur very slowly over time with population
linear cities is simply not tenable. physically relocating in response to the changing pattern
The problem with specifying a linear city in the way of accessibility that is formed as populations relocate.
proposed is that there is no sense in which the city is Given the 400 modules comprising the original sequence,
in equilibrium. If we array individual units in such a
sequence, their relative positioning makes a difference
to how close they are to one another. The central seed 2
In fact there are 401 cells or zones in the city arranged around one central
site x0 is the most accessible using the formula in eq. seed cell xo with two linear arms left L as xL200, xL199, xL198, …, xL1, to the cen-
(4) as we have shown in Fig. 4 but the distribution of tral seed xo and then from the seed to the right as xR1, xR1, xR3, …, xR200.
Batty Computational Urban Science (2022) 2:8 Page 9 of 17

Fig. 5 Agglomeration to a More City-Like Form. The graphs are profiles across the 401 cells on the horizontal axis where the seed site is the central
site. The vertical axis is population scaled to sum to the total population which for each example is 401 units

with some 2400 persons in each neighbourhood module, – land and distance – that can be easily related to their
then it might be possible to demonstrate a slow move- demand (Henderson and Quandt, 1958).
ment of population from these modules that could be fit-
ted to a feasible evolution of such an equilibrium, even 4 The polycentric Linear City
perhaps in a real city. In fact the differences between Linear cities usually have one central seed site and
each iteration with respect to populations in the series although historically, cities tend to only have one core,
of cells, fall very rapidly at first and within 10 iterations since the industrial revolution as cities began to grow
to all intents and purposes, the new population distribu- and new technologies enable rapid movement to distant
tion measured by ∑s|Ps(t + 1) − Ps(t)|/Ps(t) is within 1% places, a degree of polycentricity has emerged. Many cit-
of its previous value for each of the four functions shown ies have begun to fuse into one another with multiple
in Fig. 5. There is little point in iterating these functions centres occurring within urban agglomerations manifest-
much further for after about 25 iterations to all intents ing a hierarchy of centres of different sizes. These sizes
and purposes the equilibrium is reached. It is worth relat- are likely to be conditioned by some economic principle
ing the way we define population and distance to accessi- of agglomeration such as that defined in central place the-
bility through other attributes of location that define the ory but here we will simply choose a relatively arbitrary
urban economy, specifically the rent payable and travel hierarchical subdivision of the line and choose differ-
cost incurred. We can relate accessibility An to rent and ent sizes of subcentre. Let us adopt a binary hierarchical
distance to transport cost d0xn and note that the utility subdivision where we begin with 401 cells and around
that is gained at each point can be defined as An d0xn. This the central cell, divide the remaining 200 cells either
relates our analysis to maximising utility where rent pay- side into two sequences of 100 in the following manner.
able and transport cost can be viewed as economic goods At the top of the hierarchy, we first divide the 400 cells
into 2. Then we divide the 200 cells in each branch of the
Batty Computational Urban Science (2022) 2:8 Page 10 of 17

Fig. 6 Multiple Centres and Edge City Effects. The black lines show the distribution of population as arranged in a hierarchy in 6(a) and as edge
effects in 6(b) and 6(c). The red lines show what the ultimate outcomes for each of these three cases as the system moves towards equilibrium.

hierarchy into two sequences of 100. We continue in this as the attractions remain relative to one another. This is
fashion and at each stage in the hierarchy, we compute easy to see for the algorithm that enables the equilibrium
the number of new subcentres as n(h) = ­2h, h = 0, 1, 2, 3, to be computed in eq. (6) is independent of scale, that is,
…. which follows the sequence of numbers of subcen- their relativity is preserved no matter how population is
tres n(0) = ­20 = 1 and then 2, 4, 8, 15, 32 = n(5) = ­25. The scaled.
total number of centres across the line is computed from When we run the model, we find it impossible to break
m(h) = ­2h + 1 − 1. We show this sequence in Fig. 6(a) not- the intrinsic symmetry of the line around its seed. The
ing that we can increase the density of polycentres by power of the central site is so great that even when we
moving to finer subdivisions at each level, using larger space significant populations at the different scales of
generative factors such as n(h) = ­3h, n(h) = ­4h and so on. hierarchy, we cannot break the agglomerative forces that
Using this algorithm, the polycentres are shown at move populations towards the central seed site. In fact
equal spacing of their subdivisions in Fig. 6. We have the hierarchical distribution around the central seed site
also specified the hierarchy down to level h = 4 where we reinforces the larger concentration towards the central
have arbitrarily specified the size of each of these cen- site and it is not surprising that the picture that emerges
tres so that they add to 250 units. In fact there are some wipes out the effects of the polycentres other than the
31 units that contain these giving an average of about 8 central seed. In fact if we position the population towards
households for the polycentres. The remaining 469 units the left and right edges of the line, that is at xL200 and at
need some level of population and thus we assume that xR200 which in terms of the list of sites are x1 and xN + 1,
each of these modules has 150/469 units which is about we still find it difficult to break the focus on x0. In fact
0.31 units per cell. As these values are all relative and only we can just about do this by pumping these populations/
used to show the attraction of populations, then we do attractions at the extreme ends of the line ever higher
not need to worry about the nature of the units as long but in essence, the initial geometry is hard to fracture
Batty Computational Urban Science (2022) 2:8 Page 11 of 17

with respect to the accessibility of the central point. In zones or locations any of which can interact with any
Figs. 6(b) and (c), we look at two cases where we arrange other. What we should now do is rework the logic of the
the population so that it increases as either end of the last section computing accessibilities for each of these
line is approached and show the ultimate distribution locations and this would require us to initiate some
in equilibrium. Despite this demonstration of the power 160,8012 = 2,585,696,160 calculations several times
of geometry in cities, it also illustrates how hard it is to over. It is impossible to run the model at this scale for it
depart from systems such as cities which grow from sin- requires more than 2.5 billion operations simply to com-
gle seed sites. pute a basic accessibility, and thus for us to proceed, it is
Of course, the population is an attraction profile that very clear that we need to reduce the scale of the prob-
is likely to be correlated with many other activities. For lem. Rather than a 401 unit line we define one 20 times
example, the population attracted to each cell gives a smaller, working with a grid of 2­ 12 = 441 zones. In this
measure of population density with the central core the mode, our style is one of demonstrating an exploration of
highest and the decline in densities quite characteristic of the properties of the linear city from which we ‘assume’
real cities. The same might be assumed with respect to we can generalise. In fact the regularity of the processes
the price of land or rent which also declines in the same involved are such that we have some confidence in being
way, so in fitting such a model these would be the indica- able make such an assertion.
tors to examine. However this is but another index of how We now need to also generalise the interaction model
infeasible such a linear structure might be in that it takes and the equilibrium conditions to the square grid. We
no account of variable demand, different preferences for will define an augmented spatial model based on a
urban living and so on. These are all features that come gamma-like distribution function which has two deter-
to bear on how such a city were to actually develop if the rence (distance) components. This can be stated as
developers at some point were to take off the constraints
and regulations on how land could be developed and how
β
people might travel. Page (1999) develops a hypotheti- frs ∝ drs exp(γ drs ) (7)
cal geometric city form whose rules establish different
where β and γ can be tuned to reflect the interaction
economic equilibria not dissimilar from the process we
between centralising and decentralising effects. There are
use here. There are other models consistent with urban
four possibilities: first if β < 0 and γ < 0, then the effect is
economic theory that focus on how individuals locate to
to heavily concentrate development close to its original
maximise their utilities in different geometrical contexts
source and this would keep the focus on the line rela-
but these tend to start with micro-economic theory and
tively intact. Second if β > 0 and γ > 0, then this would
then show how the equivalent model can be distorted
push development to the edge of the system, away from
with different geometries rather than starting with differ-
its original source. With the other two combinations
ent geometries and building equivalent models. This rep-
β > 0 and γ < 0 and β < 0 and γ > 0, it is not possible to
resents another dimension to this work where we refer
guess the balance of forces in advance. However these
the reader to Solow and Vickery (1971) and Wang (1993).
forces would operate relentlessly from the initial condi-
tions where all unit populations located on the line. In
5 Transforming one‑dimension into two the first time interval t = 1, there are 420 empty locations
So far we have kept the intrinsic linearity of urban form to be considered and there are different ways these might
and have simply shown how agglomeration economies be populated once the initial clusters on the line are given
are simulated by computing the accessibility at differ- the opportunity to move as they search for locations that
ent points along the line. We began with each point on optimise their accessibility.
the line weighted equally in terms of its contribution to Following eq. (6), the equilibrating process can be
the accessibility and it is the relative distances between stated again as
each point and all others to which it can be linked that
determine its actual accessibility. There are several dif-   
ferent ways of keeping the city on the line and deter- Ps (t + 1) = P
r
α
Ps (t)drs exp(γ drs )/ α
Ps (t)drs
s
exp(γ drs )
mining the population density of each point according (8)
to different forms of accessibility as we have just illus- where we now weight the gamma-like function with
trated but to demonstrate how the city may radically the population Ps(t) from the previous time period. At
depart from the line, we need to consider all the poten- the beginning of the process, the population in the loca-
tial space around the line. The easiest way to do this is to tions off the line are all zero and in this first time inter-
consider the new city area as a square grid 401 units or val, we need to enable these by excluding the measure of
clusters by 401 which generates a total of 4 ­ 012 = 160,801
Batty Computational Urban Science (2022) 2:8 Page 12 of 17

Fig. 7 Spreading the City on a Line into the Two-Dimensional Space. a) The original line left, and right b) temporal colouring, starting with the line
spreading out, but retaining a measure of linearity

population from the distribution function. Therefore we 7(a) as the starting point for this and subsequent vari-
set this as ants while in 7(b), the pattern is coloured by time across
the spectrum from red-brown to blue.
   We can easily adapt this process to scaling the loca-
α α
Ps (1) = P drs exp(γ drs )/ drs exp(γ drs ) (9) tions according to the level of accessibility at each loca-
r s
tion. We scale the population (which we assume as the
and using Ps(t = 1), we then initiate the process of itera- total number of units (441)) and in Fig. 8, we show this
tion implied in eq. (8). This is likely to converge due to process where both colour is used for the temporal loca-
the fact that the process is consistent, regular and well- tion and each circle is proportional to the size of popula-
behaved and the basic spatial form is symmetric. tion in each unit. Note that in these two cases, we have
We will explore three variants for showing how move- assumed that the only deterrence is the first distance
ment towards an equilibria can be defined by maximis- effect and we thus set β =  − 2 with γ = 0. This is in effect
ing local accessibilities. The first variant is where the the inverse square law but of course we could use any
locations defined on the original line, do not change variant of the function in eq. (7). We have explored some
in their population but that at each time period, a new of these more casually and they do not produce any real
unit of population is located in the empty cell that has differences from the outcomes that we show in this sec-
the highest accessibility. This occurs one at a time and tion. We should also note that the order in which loca-
once the location is populated, it cannot be repopulated tions are considered occurs according to the labelling of
or lose population. So this process starts with the origi- cells and this does occasionally impose some asymmetry
nal 21 units on the line across the region, and then each on the form. In fact this does occur in Fig. 7 where the
of the remaining 420 locations are populated in order of horizontal line across the square grid continues to be
their highest accessibilities. Ties are broken arbitrarily reflected in the ultimate spread. But this is not so in Fig. 8
but the final total population is ultimately the number where most of the asymmetry that we can clearly observe
of cells in the space, namely 441. As the process occurs as the sequence of iterations moves towards an equilib-
one cell at a time, the space fills up and as the time t rium, is ironed over the 100 iterations used to show that
reflects the strict order of computation, we can colour an equilibrium occurs.
the sequence of cells coming on stream according to this We will now show how different values of the param-
variable. In Fig. 7, we show this sequence and this pro- eters of the function frs in eq. (7) control the spread into
vides a catalogue as to what cells get developed where the space. The default model that we again use is where
and when. You can see the city spreading out and ulti- β =  − 2 with γ = 0 (the traditional inverse square law of
mately becoming ever more monocentric around the distribution) and in Fig. 9, we show the first ten itera-
centre of the space which is the centre of the line. The tions. This shows how the line begins to spread out into
original line is still part of the pattern but it gradually a form that ultimately becomes circular and symmetric.
reduces in strength. The units on the line are shown in In fact the ultimate spread which is symmetric within the
Batty Computational Urban Science (2022) 2:8 Page 13 of 17

Fig. 8 Spreading the City and Scaling Population into the Two-Dimensional Space

Fig. 9 Convergence of the Model Function frs ∝ d−2

rs Used to Spread The Population on the Line

grid occurs quite rapidly which is close to the final out- 6 Different initial configurations based on the line
come after 4 iterations and by the 20th is more or less Although our focus here is on linear cities, we can use
unchanging although we have taken the number of itera- the logic of maximising local accessibilities to explore
tions to the point where t = 100. many different initial configurations and how their
We can now illustrate the nine possible solutions at forms are transformed to a stable equilibrium. We will
t = 100 defined by the limits of the parameters, set at all define three simple extensions here which consist of
combinations of the values from [β =  − 5, β = 0, β = 5] first adding another line which crosses the original line
and [γ =  − 1, γ = 0, γ > 1]. The nine patterns are shown in at the centre of the grid defining the overall city space.
Fig. 10 where we have arrayed them in 3 by 3 formation We then can add four locations in the middle of each of
and this is a clear illustration of the effects of the central- the four squares that define the quadrants of the cross
ising and decentralising forces that originate from every- and then we can complete this structure by adding pop-
body living on the line in the first instance. ulation locations at the four extreme points defining the
Batty Computational Urban Science (2022) 2:8 Page 14 of 17

β
Fig. 10 The Limits of the Solution Space for Different Parameter Values of drs exp (γdrs )

corners of the square grid. The first set of experiments extent, the limitation of this analysis is that we are not
involve using the standard inverse square model with able to dimension any real time to this process for any
β =  − 2 with γ = 0 and iterating each of the three new process of change is much more lumpy, and thus likely
configurations which we illustrate in the first column of to take a very long time to converge to a stable state.
Fig. 11. In this figure, we show each initial configura- We can also explore what happens when the parame-
tion from t = 0 through the first iterations from t = 0, 1, ters β and γ are varied across the range that we illustrated
2, 3, … to the final iteration which marks a strong con- in Fig. 10. When the values of these parameters are posi-
vergence where t = 300. tive and large, then the initial population will be rapidly
From this figure, it is clear that the original form diffused to the edge of the city space, defining a process
has little impact on the final configurations which all of decentralisation. When these parameter values are
appear similar. In short, the original geometry is grad- negative and large, then this mirrors a process of centrali-
ually ironed out, smoothed out if you like, from the sation but because the process in general is one of diffu-
initial city space. In fact we can compare this to the sion across the space from a dominant central location,
original configuration of the simplest line in Fig. 9 and the long term equilibrium still reflects the central loca-
the final outcome appears very similar to that shown tion but this is smoothed out. We show these outcomes
above in Fig. 11. To an extent, this is a very obvious for the three different configurations in Fig. 12 where it is
result in that the process of iteration to a new equilib- clear that the initial configurations make little difference
rium is akin to a diffusion process where in the limit to the long term structures and in comparison with the
everything influences everything else and the original outcomes in Fig. 10, the decentralised patterns are pretty
distribution is gradually lost from the picture. We have similar to one another whereas the centralised ones differ
continued this iteration for many iterations and com- somewhat. There is much work still to do on exploring
puted differences between each successive pattern. This the actual equilibrium properties of these solutions.
does indeed converge very rapidly to a stable form but We have computed many measures of spread and den-
this form is not an even distribution across the space sity for these different solutions and we will show these
for the original geometry and the order of computa- in a future extension of the ideas in this paper. The distri-
tion still make a difference to the ultimate form. To an butions of population can also be plotted and across this
array of forms, we can detect some of these pertaining
Batty Computational Urban Science (2022) 2:8 Page 15 of 17

Fig. 11 The Convergence of Configurational Variants of the Linear City

to the frequencies of the size distributions we observe in Neom and although we have not specified a temporal
real cities. This provides us with an additional filter that dimension on how this city on a line might develop into
we can use to suggest which simulations are closer to real a more concentric form, our illustrations in the previous
cities, and we also need to explore in much more detail sections do suggest that this might happen quite quickly.
the values of the parameters that influence the accessibil- Once the line is established, populations could then begin
ities and the trip patterns that lie behind these different to operate a market for land which would reinforce issues
forms. We have computed many of these already but with of agglomeration and accessibility spreading new growth
all work in progress, we have yet to report results. in the city ultimately across the entire space.
In fact, prior to the first industrial revolutions, many
7 Conclusions: how real is this analysis? smaller settlements where walking was the dominant
There have been many suggestions that the ideal city of mode of transportation complemented by horse drawn
the future should be linear in form so that accessibility to transport, were linear in form. This was particularly true
the countryside and to others spread out along the line of small villages. It took new transportation technologies
can avail themselves of fast transportation. In fact, the to change this and in some respects, it might be argued
linear city pays little regard to the total energy used and that several of the proposals for future cities based on lin-
to questions of how fast transportation can be accessed. ear forms represents a flash back to an earlier age where
Moreover the sheer level of control required to imple- transportation was of a different nature but also to an
ment such structures is likely to be almost impossible to age when new transportation technologies emphasised
achieve as and this is very clear from the fact that the idea the notion of travelling in a straight line. Only when one
has never resulted in truly linear cities continuing to exist considers the behaviour of many individual agents all
in their original pristine form. As in Brasilia, La Cuidad competing with each other for the most accessible space
Lineal, in Stalingrad and many more applications, once does the linear city idea appear to be inconsistent with
control is relaxed, these cities begin to grow and develop present patterns of human movement behaviours. There
in more traditional ways based on messier, cluster-like is still much to be done on taking these ideas forward
radially concentric forms. The same fate probably awaits for the model we have proposed is highly unrealistic in
Batty Computational Urban Science (2022) 2:8 Page 16 of 17

Fig. 12 Centralised & Decentralised Equilibria for Three Elaborations on the Linear City

many respects and can certainly be improved. However Author’s contributions

I am the sole author, wrote the paper, produced the research, wrote the com-
we do not consider that the results of such extensions puter programs and generated the graphics. The author read and approved
which involve adding much more richness and random- the final manuscript.
ness would radically change the conclusions that we
Funding
have come to here. In this sense, our conclusion is that This research is supported by The Alan Turing Institute under the QUANT2–
the linear city will remain a fiction, a speculation, more Contract–CID–3815811, and the UK Regions Digital Research Facility (UK
a thought experiment than a potential reality, but some- RDRF) EP/M023583/1.
thing that is not without value in stretching our imagi- Availability of data and materials
nations about possible urban futures. Even now in Neom, All data is simulated within the model and does not need any copyright
there is a proposal for building a virtual reality prior to permissions.
anyone even living in the city, a digital twin that exists in
the metaverse (Bell, 2021) on which one can experiment Declarations
and explore the sorts of different futures that we have Competing interests
presented here. There are no competing interests.

Acknowledgements Received: 7 February 2022 Accepted: 8 February 2022

There are no specific acknowledgements to other researchers.
Batty Computational Urban Science (2022) 2:8 Page 17 of 17

References
Antyufeev, A. V. & Antyufeeva, O. A. (2019). Linear cities: Controversies, chal-
lenges and prospects. IOP Conference Series: Material Sciences and Engi-
neering. 687, 1–6. doi:https://​doi.​org/​10.​1088/​1757-​899X/​687/5/​055025.
Banerji, R. (2012). Niemeyer’s Brasilia: Does it work as a city? BBC world service. 7
December 2012. https://​www.​bbc.​co.​uk/​news/​magaz​ine-​20632​277
Bell, J. (2021). Saudi’s NEOM will have a digital twin in the ‘Metaverse’. al-
Arabiya news. https://​engli​sh.​alara​biya.​net/​News/​gulf/​2021/​10/​27/​Saudi-
s-​NEOM-​will-​have-a-​digit​al-​twin-​in-​the-​metav​erse-​FII
Boileau, I. (1959). La ciudad lineal: A critical study of the linear suburb of
Madrid. The Town Planning Review., 30, 230–238.
Clarke, K. C. (2013). What is the world’s oldest map? The Cartographic Journal.,
50, 136–143.
Collins, G.R (1959) Linear planning throughout the world. Journal of the Society
of Architectural Historians. 18, 74-93. Reprinted in Ekistics. 1960, 9, 242–25.
Doxiadis, C. A. (1967). On linear cities. The Town Planning Review., 38, 35–42.
Furundzic, D. S., & Furundzic, B. S. (2013). Infrastructure corridor as linear city.
University of Belgrade, Faculty of Architecture, Bulevar Kralja Aleksandra
73, Belgrade, Serbia.
Gold, J. R. (1995). The MARS plans for London, 1933-1942: Plurality and experi-
mentation in the city plans of the early British modern movement. The
Town Planning Review., 66, 243–267.
Gordon, A. (2021). Saudi Arabia resurrects discredited "linear city" concept
in bigger, dumber form. Vice Media Group. https://​www.​vice.​com/​en/​
artic​le/​m7a598/​saudi-​arabia-​resur​rects-​discr​edited-​linear-​city-​conce​
pt-​in-​bigger-​dumber-​form.
Hanson, I. A. (2020). Hyperloop transport technology assessment and system
analysis. Transportation Planning and Technology., 43, 8803–8820.
Henderson, J. M., & Quandt, R. E. (1958). Microeconomic theory: A mathematical
approach. McGraw-Hill.
Hillier, W. J. (1996). Space is the machine. Cambridge University Press.
Hodder, I. (2005). Women and men at Çatalhöyük. Scientific American. Special
Editions, 15, 34–41.
Jacobs, J. (1961). The death and life of great American cities. Random House.
March, L. (1967a). Let’s build in lines. New Society. 20 July, 75–77.
March, L. (1967b). Homes beyond the fringe. Architects Journal. Royal Institute
of British Architects, August, 334–337.
Marshall, S. (2005). Streets and patterns. Routledge.
Milyutin, N. A. (1930, 1975). Sotsgorod: The problem of building socialist cities.
The MIT Press.
Page, S. E. (1999). On the emergence of cities. Journal of Urban Economics., 45,
184–208.
Solow, R. M., & Vickery, W. S. (1971). Land use in a long narrow city. Journal of
Economic Theory., 3, 430–447.
Tobler, W. R. (1970). A computer movie simulating urban growth in the Detroit
region. Economic Geography, 46, 234–240.
Wang, P. (1993). Agglomeration in a linear city with heterogeneous house-
holds. Regional Science & Urban Economics., 23, 291–306.
Wiebenson, D. (1960). Utopian aspects of Tony Garnier’s “Cité Industrielle”.
Journal of the Society of Architectural Historians., 19, 16–24. https://​doi.​org/​
10.​2307/​987962
Worth, R. F. (2021). The dark reality behind Saudi Arabia’s utopian dreams. New
York Times. February 24, 2021. https://​www.​nytim​es.​com/​2021/​01/​28/​
magaz​ine/​saudi-​arabia-​neom-​the-​line.​html

Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in pub-
lished maps and institutional affiliations.