• No results found

Visualizing the cognitive-cultural : a longitudinal spatial analysis of municipal business registration data from 2004, 2008, and 2012

N/A
N/A
Protected

Academic year: 2021

Share "Visualizing the cognitive-cultural : a longitudinal spatial analysis of municipal business registration data from 2004, 2008, and 2012"

Copied!
159
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Visualizing the Cognitive-Cultural

Economy of Amsterdam

A Longitudinal Spatial Analysis of

Municipal Business Registration Data

from 2004, 2008, and 2012

Lucas C. Reckhaus

August 2016

(2)

I University of Amsterdam, 2016

Graduate School of Social Sciences Human Geography: Urban Geography Master's Thesis, MSc

By: Lucas C. Reckhaus Student Number: 11127767

Supervisor: Dr. Robert Kloosterman Second Reader: Dr. Wouter van Gent

Abstract

The cognitive-cultural economy (CCE) is growing as an academic concept to describe the new economy of cities at the outset of the 21st century. It is considered to primarily consist of business and financial services, hi-tech, and creative industries. A gap in the literature has been identified concerning the visualization of this economy as it is spatially located across a given area. Doing so provides distinct advantages to understanding its role in the city. Through a longitudinal time-sequence spatial analysis— using municipal business registration data—this thesis plots the course of the CCE in Amsterdam in the years 2004, 2008 and 2012, deconstructing where in the city the three sectors locate, to what extent they overlap, and an exploration of influencing factors as to why they may locate in a given place.

(3)

II Für Hermann

(4)

III

Foreword

“From now on, i'll describe the cities to you,” the Khan had said, “in your journeys you will see if they exist.”

But the cities visited by Marco Polo were always different from those thought of by the emperor. “And yet I have constructed in my mind a model city from which all possible cities can be deduced,” Kublai said. “It contains everything corresponding to the norm. Since the cities that exist diverge in varying degree from the norm, I need only foresee the exceptions to the norm and calculate the most probable combinations.”

“I have also thought of a model city from which I deduce all the others,” Marco answered. “It is a city made only of exceptions, exclusions, incongruities, contradictions. If such a city is the most improbable, by reducing the number of abnormal elements, we increase the probability that the city really exists. So I have only to subtract exceptions from my model, and in whatever direction I proceed, I will arrive at one of the cities which, always as an exception, exist. But I cannot force my operation beyond a certain limit: I would achieve cities too probable to be real.”

(5)

IV

Table of Contents

Abstract ... I Foreword ... III Table of Contents ... IV List of Figures ... IX List of Abbreviations ... X 1) Introduction ... 1

1.1 The Importance of Data Visualization ... 1

1.2 Scott's Cognitive-Cultural Economy in a Nutshell ... 2

1.3 Unpacking the CCE: Finding the Common Thread ... 3

1.4 The Role of the Global City ... 4

1.5 The Use of Maps in This Research ... 5

2) Research Question ... 6

3) Theoretical Framework ... 7

3.1 Defining the Cognitive-Cultural Economy: Who Belongs and Who Does Not? ... 7

3.2 Main debates in the role of data Visualization ... 9

3.3 Clustering: Of What, Where, Why and How? ... 12

3.4 Global Linkages in Amsterdam and their impact on the CCE ... 17

3.5 Conceptual Model ... 21

4) Methodology ... 22

4.1 Case Set ... 22

4.2 Specific Case ... 23

4.3 Sub-Cases ... 24

4.4 Type of Case Study ... 25

4.5 Relationship between Cases & Variables ... 25

(6)

V

4.7 Inferences and Predictions ... 27

5) Operationalization: Data Selection and Mapping Methods ... 28

5.1 Business Registration Data: Years Covered, Information Stored, Locational Attributes ... 28

5.2 Selecting the Amsterdam Cognitive-Cultural Economy from Government SBI Codes ... 30

5.3 Geocoding Data and Creating Choropleth Maps ... 31

5.4 Isolating Neighborhoods with the Highest CCE Activity ... 34

5.5 Cluster Mapping with 25 meter buffers ... 36

5.6 Getis-Ord Gi* Statistical Hot Spot Method ... 37

5.7 Measurement Scale: Map Legends, Standard Deviations, Sum Counts, Color Choices .... 38

6) Data Analysis: Tracing the Changes of the Cognitive-Cultural Economy in Amsterdam ... 43

6.1 Overall Changes in Composition from 2004 through 2012 ... 43

6.2 Business and Finance ... 46

6.2.1Business and Finance in 2004 ... 46

6.2.1.1 Citywide Patterns of Business and Finance in 2004 (Locational Density and Pattern) 46 6.2.1.2 Area Specific Patterns of Business and Finance in 2004 ... 47

6.2.2 Business and Finance in 2008 ... 48

6.2.2.1 Citywide Patterns of Business and Finance in 2008 (Locational Density and Pattern) 48 6.2.2.2 Area Specific Patterns of Business and Finance in 2008 (Hot Spot Clusters) ... 49

6.2.3 Business and Finance in 2012 ... 51

6.2.3.1 Citywide Patterns of Business and Finance in 2012 (Locational Density and Pattern) 51 6.2.3.2 Area Specific Patterns of Business and Finance in 2012 (Hot Spot Clusters) ... 52

6.3 Creative Industries ... 54

6.3.1 Creative Industries in 2004 ... 54

6.3.1.1 Citywide Patterns of Creative Industries in 2004 (Locational Density and Pattern) .... 54

6.3.1.2 Area Specific Patterns of Creative Industries in 2004 (Hot Spot Clusters) ... 55

6.3.2 Creative Industries in 2008 ... 58

6.3.2.1 Citywide Patterns of Creative Industries in 2008 (Locational Density and Pattern) .... 58

(7)

VI

6.3.3 Creative Industries in 2012 ... 61

6.3.3.1 Citywide Patterns of Creative Industries in 2012 (Locational Density and Pattern) .... 61

6.3.3.2 Area Specific Patterns of Creative Industries in 2012 (Hot Spot Clusters) ... 62

6.4 Hi-Tech ... 66

6.4.1 Hi-Tech in 2004 ... 66

6.4.1.1 Citywide Patterns of Hi-Tech in 2004 (Locational Density and Pattern) ... 66

6.4.1.2 Area Specific Patterns of Hi-Tech in 2004 (Hot Spot Clusters) ... 67

6.4.2 Hi-Tech in 2008 ... 69

6.4.2.1 Citywide Patterns of Hi-Tech in 2008 (Locational Density and Pattern) ... 69

6.4.2.2 Area Specific Patterns of Hi-Tech in 2008 (Hot Spot Clusters) ... 70

6.4.3 Hi-Tech in 2012 ... 72

6.4.3.1 Citywide Patterns of Hi-Tech in 2012 (Locational Density and Pattern) ... 72

6.4.3.2 Area Specific Patterns of Hi-Tech in 2012 (Hot Spot Clusters) ... 73

6.5 Areas of Multi-Functional Use (Multi-Sector Hot Spots) ... 75

6.5.1 Multi-Functional Spaces in 2004 ... 76

6.5.2 Multi-Functional Spaces in 2008 ... 78

6.5.3 Multi-Functional Spaces in 2012 ... 80

7) Interpretation of Results ... 83

7.1 Sector-Specific Agglomerations of Hot Spot Clusters & Ellipses Movement ... 83

7.1.1 Business and Finance ... 83

7.1.2 Creative Industries ... 85

7.1.3 Hi-Tech ... 86

7.2 Inter-sectoral Agglomerations of Clusters ... 87

7.3 Neighborhood Characteristics of Sector Activity ... 87

7.4 Availability and Cost of Spaces ... 89

7.5 The Role of the 2008 Financial Crisis ... 91

7.6 Impact of 1-Person Firms ... 92

(8)

VII

8.1 Answering the Research Question ... 93

8.2 Possible directions for further research ... 96

References ... 98

Appendix A, B, C ... 104

A Summary Statistics ... 104

A1 Registered Businesses in Amsterdam and Cognitive Cultural Economy Entries ... 104

A2 Total Number of Cognitive Cultural Economy Sector Entries Per Year ... 104

A3 Ghost Firm Data ... 105

A4 Information on 1-Person Firms ... 105

A5 Getis-Ord Gi* Statistical Hot Spot Method Cluster Details ... 106

A6 Informational Attributes ... 107

B Occupational Categories Selected and SBI Codes ... 108

B1 Business and Finance ... 108

B2 Creative Industries ... 110

B3 Hi-Tech ... 111

C Maps ... 115

C1.1 Business & Finance 2004 (Neighborhood Composition) ... 115

C1.2 Business & Finance 2004 (Top Neighborhoods) ... 116

C1.3 Business & Finance 2004 (Hot Spot Clusters) ... 117

C1.4 Business & Finance 2008 (Neighborhood Composition) ... 118

C1.5 Business & Finance 2008 (Top Neighborhoods) ... 119

C1.6 Business & Finance 2008 (Hot Spot Clusters) ... 120

C1.7 Business & Finance 2012 (Neighborhood Composition) ... 121

C1.8 Business & Finance 2012 (Top Neighborhoods) ... 122

C1.9 Business & Finance 2012 (Hot Spot Clusters) ... 123

C1.10 Business & Finance, Directional Distribution (04, 08, 12) ... 124

C2.1 Creative Industries 2004 (Neighborhood Composition) ... 125

C2.2 Creative Industries 2004 (Top Neighborhoods) ... 126

(9)

VIII

C2.4 Creative Industries 2008 (Neighborhood Composition) ... 128

C2.5 Creative Industries 2008 (Top Neighborhoods) ... 129

C2.6 Creative Industries 2008 (Hot Spot Clusters) ... 130

C2.7 Creative Industries 2012 (Neighborhood Composition) ... 131

C2.8 Creative Industries 2012 (Top Neighborhoods) ... 132

C2.9 Creative Industries 2012 (Hot Spot Clusters) ... 133

C2.10 Creative Industries Directional Distribution (04, 08, 12) ... 134

C3.1 Hi-Tech 2004 (Neighborhood Composition) ... 135

C3.2 Hi-Tech 2004 (Top Neighborhoods) ... 136

C3.3 Hi-Tech 2004 (Hot Spot Clusters) ... 137

C3.4 Hi-Tech 2008 (Neighborhood Composition) ... 138

C3.5 Hi-Tech 2008 (Top Neighborhoods) ... 139

C3.6 Hi-Tech 2008 (Hot Spot Clusters) ... 140

C3.7 Hi-Tech 2012 (Neighborhood Composition) ... 141

C3.8 Hi-Tech 2012 (Top Neighborhoods) ... 142

C3.9 Hi-Tech 2012 (Hot Spot Clusters) ... 143

C3.10 Hi-Tech Directional Distribution (04, 08, 12) ... 144

C4.1 Multi-Functional Cognitive-Cultural Areas 2004 ... 145

C4.2 Multi-Functional Cognitive-Cultural Areas 2008 ... 146

(10)

IX

List of Figures

Figure 1: Space Types and their Use for Work………..…….……….……….….. 13

Figure 2: Sources of Locational Competitive Advantage. ……….……….……….………...……14

Figure 3: Four Ions of Economic Geography..……….………….…….………...……..14

Figure 4: The Structure and Dynamics of Local Buzz and Global Pipelines…….……….…..…..15

Figure 5: The Extended Workplace Cluster...………..………..…..……16

Figure 6: Storper’s ‘holy trinity’………..…...…….16

Figure 7: Atlas 1-Flow Lines [Amsterdam]: Global Linkages………...……….……..……..17

Figure 8: Atlas 2-City Views [Amsterdam]: Regional Compacting………..….…….18

Figure 9: Alpha (α) Cities………..……….…….19

Figure 10: The World of Alpha (α) Cities………..….….……20

Figure 11: Shape of Connectivity amongst World Cities of Same Tier….……….…...…...……..22

Figure 12: Neighborhood Boundaries of the Canal District……….………….….…….………....24

Figure 13: Europe GDP 2007 & 2009………..………….………..44

Figure 14: Amsterdam Neighbourhood Overview……….………...…………88

Figure 15: Vacant Buildings (non-residential)……….…….….………..89

Figure 16: Relative Housing Prices in Amsterdam 2010……..……….………..….…….………..90

Figure 17:Cultural-historic values in Amsterdam………..………..….………..…….94

(11)

X

List of Abbreviations

47 (2-digit) Amsterdam Ward Code

47c (2-digit and 1-letter) Amsterdam Neighborhood Code α: Alpha (1st Tier)

β: Beta (2nd Tier)

ɣ: Gamma (3rd Tier)

ArcMap 10.1: Environmental Systems Research Institute (ESRI) Mapping Software A10 Ring: Primary Highway Loop Around Amsterdam

BF: Business and Finance CI: Creative Industries

CBS: Centraal Bureau voor de Statiestiek (Central Bureau of Statistics of the Netherlands) CCE: Cognitive Cultural Economy

GaWC: Globalization and World Cities Research Network GIS: Geographic Information System

OIS: Onderzoek, Informatie en Statistiek (Amsterdam Municipal Office of Research, Information and Statistics)

SBI: Standaard Bedrijfsindeling (Standard Industrial Classification Code) SD: Standard Deviation

(12)

1

1) Introduction

1.1 The Importance of Data Visualization

Humans are foremost visual beings. The enduring importance of art over millennia perhaps best epitomizes the fundamental importance of visualization as a primary method to absorb information and facilitate its transmission to others. However the importance of visualization as a tool for learning is also widely substantiated in pedagogic research (Rasul et al., 2011; Agostinho, 2011). At the dawn of the digital age, it is more important than ever to reaffirm the central importance of visualization as a tool for learning and storytelling. This is because the digital revolution continues to engender the largest ever collection of information in human history, as every scant piece of information is slowly but surely becoming digitally processed and stored (Wyly, 2013: 392). What to do with this new wealth of information is a great challenge. How does one draw out the pieces of meaning and relevance? How does one take data and tell a story from it? These are fundamental problems that governments, private entities, and academic researchers are struggling with (Goldstone, 2015: 249). For data does not speak on its own—it always needs to be interpreted, and the results synthesized in a way that can be easily conveyed to others. In conveying meaning to others data visualization stands out as exceptionally effective.

While there are a variety of methods to data analysis, and some naturally suit certain types of data better than others, data visualization—through the use of charts, maps, webs, and other spatial forms—is an irreplaceable tool. This thesis takes this premise and, through the use of maps, applies it to a specific case, the cognitive-cultural economy (CCE) of Amsterdam, and specific data set, the city's business registration database.

(13)

2

This introduction begins by: defining the cognitive-cultural economy; deconstructing the most contentious aspects of the theory; highlighting the theory that underlines data visualization; and the way in which spatial mapping can be applied to the case of Amsterdam's CCE to advance the understanding of its distribution and composition within the city.

1.2 Scott's Cognitive-Cultural Economy in a Nutshell

Allen Scott in his (2012) work A World In Emergence provides the following definition of the cognitive-cultural economy, in which he argues the CCE forms a crucial part of the new economy of the 21st century:

I prefer to substitute the term cognitive-cultural economy for the once useful but now essentially obsolete label “postfordism.” The new division of labor that is being brought forth as this economy forges ahead can be identified at the outset in terms of an upper fraction of well-paid, highly qualified cognitive and cultural workers, and a lower fraction of low-wage workers, increasing numbers of them in services and most of them with minimal formal qualifications.... the upper fraction of the labor force... is nowadays increasingly called upon to exercise its personal talents and creativity in much more open-ended tasks that entail key skills such as analytical shrewdness, deductive reasoning, technical insight, interpersonal judgement, imaginative thinking, cultural sensibility, story-telling abilities and so on (p.37).

Yet this definition alone is insufficient to describe the full parameters of the new economy. More precisely, Scott claims the upper-fraction of this new economy is composed predominantly out of three specific sectors: hi-tech, business and finance, creative-industries (Scott, 2012: 41- 42). While this is a more concise definition of what constitutes the CCE, there is an inherent difficulty to this categorization as well; the term encompasses a vast range of occupations. Hartley (2005: 23) hits the nail on the head when he says of only speaking of the creative industries that they “...are so varied in scale, organization, and sector of economic activity that they are barely recognizable as a coherent

(14)

3

object of analysis within this framework. As a result, they don't always show up well in the places where industry policy is habitually discussed, in government or in business.” Key to Hartley's statement is the issue of identifying the CCE in a given setting, so that it may be more effectively understood and engaged in the arenas of government and business. Herein we find the applicability of data visualization to this case.

1.3 Unpacking the CCE: Finding the Common Thread

At this point it is necessary to ask what really unites the three CCE sectors together? Despite all of Scott's musings they are not fundamentally one and the same. For example, how are a multinational bank, a small tech firm, and an independent art gallery, somehow complementary to each other and intrinsically related industries? Instead, perhaps it is more apt to say that they are all pieces of the latest recipe which makes specifically the city a powerhouse of social and economic activity in the world today—a synthesis of businesses and financial services, technology-innovators and cultural and artistic milieus and institutions. This is the general opinion of so called “Floridians” (disciples of Richard Florida and his three T's—Technology, Talent, Tolerance) who say these qualities can elevate a city to that of a cognitive-cultural hub by providing the urban climate that attract both labor and business (Peck, 2012: 462-3). This creates an interesting dynamic, for while these three sectors may be critical ingredients to the post-modern city, their relationship is fragile—for there exists no intrinsic sense of solidarity and cohesion amidst the three CCE sectors. Such macro expressions of economic solidarity rarely materialize at the personal or small business level. Capitalism—with its unbending circular recipe for growth through the cycle of investment, innovation, production, profit and re-investment— leaves minimal room for individual employers to consider economic altruism towards others. Therefore

(15)

4

there exists a risk to speaking in such broad economic concepts as Scott and others do: that the variety of input needs between business and finance, hi-tech, and creative industries become paved over by a purported uniform set of criteria for economic growth (as is the tendency in political discourse). Therefore understanding the distinctions within Scott's cognitive-cultural economy is of intrinsic importance to promoting the CCE in an informed and equitable way. Important to measure then— because of their large explanatory potential in regards to urban trends—are the spatial distribution and composition of these three sectors over time and space in a given city. Understanding the spatial relationships between these three sectors is an important part to understanding the similarities and differences between them in an urban setting.

1.4 The Role of the Global City

This thesis takes the view that—broadly speaking—the city is the central element of the post-industrial economy (Beaverstock et al, 2000; Bathelt et al, 2004). Scott (2012:12) agrees, saying “...current rounds of capitalist development focused on the central sectors of the cognitive-cultural economy are bringing into existence a distinctive third wave of urbanization and spatial development.... the main expressions of this third wave are to be found in a network of large metropolitan areas...” However not all cities are equally embedded in the new economy. The conditions that give rise to the CCE take place overwhelmingly in Western-style global cities. New York, London, Paris, Hong-Kong, Singapore, Tokyo—all are metropoles where this economy is the new standard. At one scale down in terms of global significance, cities like Amsterdam mirror this same economy. Thisse (2010: 282) explains the fundamental structural dynamic that drives this urban configuration:

...the distinctive feature of a city is its very high density of activity, which allows agents to be close to one another. Households and firms seek spatial proximity because they

(16)

5

need to interact on a daily basis for a variety of economic and social reasons. This need has a gravitational nature in that its intensity increases with the number of agents set up nearby and decreases with the distance between locations.

The same reasoning can be extended to explain the preference of the CCE to locate in global cities, and why it can be argued that a direct relationship between business and finance, hi-tech, and the creative industries exist within the context of the city, in so far as their location in the city nourishes or detracts from their ability to attract workers, customers, advertise, and become an established part of the urban milieu (Wolfe & Gertler, 2004; Wyly, 2013; Porter, 2000; Martins, 2015).

1.5 The Use of Maps in This Research

The goal of this work then is to measure, through various mapping techniques, the spatial footprint of the cognitive-cultural economy in the city of Amsterdam over an eight year time span during the intervals of 2004, 2008 and 2012. This is done through the use of a geographic information systems program ArcMap 10.1. A variety of mapping techniques are used which together form a coherent picture of the location and composition of business and finance, hi-tech, and the creative industries. Simple color-scale choropleth maps along the lines of Amsterdam's neighborhood boundaries form the first set of maps utilized. Next, bar-chart graphics are utilized to bring greater visual emphasis to only the highest values measured. Statistical hot spot cluster mapping with the use of the Getis Ord-Gi* formula is then applied to highlight data values at an even finer scale than the neighborhood. These results are grouped into the most logical spatial agglomerations of CCE data for each sector in the city, as well as the areas of the city where the most heavy overlap occurs. Lastly a spatial statistical method is used to create elliptical visualizations which depict the spatial extent of each sector within one standard deviation of their average overall location across the city.

(17)

6

2) Research Question

How has Amsterdam's cognitive-cultural economy changed in spatial distribution from 2004 to 2012? Sub Questions:

—Are there different patterns in time and space for the three constituent sectors of the CCE (hi-tech, business and finance, creative industries) in Amsterdam?

—Are there areas of the city overlapped by activity clusters of all three CCE sectors that are therefore uniquely multi-functional to the economy in Amsterdam?

—Are there ulterior aspects of Amsterdam's built environment—such as parks, canals, and major thru-fares, social spaces, average rents—which appear to spatially relate to the clusters identified?

—What does visualizing the changing composition of the creative industries in Amsterdam over time and space reveal about the value of mapping as a tool for urban economic analysis?

(18)

7

3) Theoretical Framework

3.1 Defining the Cognitive-Cultural Economy: Who Belongs and

Who Does Not?

This rise of the CCE with its new labor prototypes and urban aesthetic represents a 'third wave' in the chronology of Western economic development. Kloosterman and Lambregts (2007: 57) sum this up saying “...we distinguish three different phases, each with its characteristic dominant form of production, its related set of agglomeration economies, and its dominant urban form. These phases are, respectively, pre-industrial (roughly 1500–1750), industrial (roughly 1750–1975), and post-industrial (1975 onward).” Cognitive-cultural theorists such as Scott claim that the new division of labor appearing in cities is the most distinguishing feature—a new demand for high levels of education in specialized fields. The effect of this has been that professional access to the CCE is increasingly exclusionary amidst a global trend of escalating costs of higher-education, along with other longstanding discriminatory trends against minorities and women in these fields (Evans, 2009: 1024; Pratt & Hutton, 2013: 94).

Therefore in the methodology of this thesis an effort has been made to define hi-tech, business and finance, and the creative industries all within the range of work that require the high education and specialization Scott describes (2012: 41-42). There are however several problems to making such categorizations, the following two of which are particularly difficult to answer.

a) What to do with trades and crafts?

One of the most problematic aspects to the definition of the CCE is its description as a high-end economy within a dichotomous city—polarized between highly-educated well-paid work, and low-wage service work. This analogy is too simplistic for all urban cases, especially as the CCE

(19)

8

increasingly becomes a panacea marketing strategy for cities (Meagher, 2013: 395-6). There is no direct relationship between having specialized skills and being well-paid for them. That relationship is context—and even more importantly—product dependent. The role of the traditional trades and crafts is a prime example of this. Consider a classic trade: the shoemaker. Mastery of this profession requires skills for which a high-degree of specialized knowledge is needed. Professionals have intimate knowledge of the materials, tools, fabrication process and possible styles that apply to their products. Individuality, creativity, and specialized knowledge are all of paramount importance. But does this qualify as a cognitive-cultural occupation? The case can be argued either way (Evans, 2009: 1013). The technical skills required would indicate “yes”. However the education needed to obtain them (usually below the university level) would indicate “no”. Therefore some argue that what really distinguishes whether a given trade or craft belongs in or out of the CCE economy is whether the product is imbued with symbolic value (Hartley, 2005: 26-27; Pratt & Hutton, 2012: 87; Kloosterman 2015: 383). A local shoe-maker repairing worn-out soles and heels is not making a fortune (normally). They represent the epitome of what the European urban working class used to constitute. On the other hand, a shoe-maker can sell luxury bespoke shoes for several hundred Euro a pair. That work is intricately tied to design, fashion, material choices and treatment processes, that give a product added value above its utility. The same applies to professions such as carpenters, tailors, and welders (among others). What is their place in the polarized world that Scott speaks of? At what point are they full production and consumption members of the CCE? Is formal higher-education really the only metric to determining entry to the CCE? Why does a design-house fit more easily into this typology than an independent artisan? These are difficult questions to answer, and very much dependent on subjective choices.

b) What about the all independent artists?

Another problem has to do with the issue of data collective and treatment. The data set used for this research is one in which the businesses of Amsterdam are registered and from which they can be

(20)

9

geographically located. But what of those individuals who work free-lance, part time, or on private commission? In the arts such work is especially common: a painter likely does not have a registered business but rents a studio to produce his/her works and then sell them at a third-party gallery (if they are successful); actors regularly switch employment between different companies, theaters, and performance spaces; musicians work on a venue by venue basis. Only a select few in these industries have the luxury of a permanent work address, such as with an established artistic institution (the National Opera for example), or at a permanent studio or performance space. Therefore in regards to this data set it must be asked just how much of the arts economy is really being captured in the category of creative-industries. Even when the Municipal government has a listing for actor or painter, it cannot be a complete index of all those that make up that economic milieu. This is an important discrepancy to acknowledge and it speaks to the intrinsic difficulty of the overall characterization of the cognitive-cultural economy.

3.2 Main debates in the role of data Visualization

The importance of visual aids in facilitating cognitive learning is a central point of emphasis in this thesis. This claim rests upon the findings of longstanding pedagogical research into the cognitive and psychological mechanisms of learning. Getman, writing in 1981 during the infancy of the digital age, aptly says “There is also more and more evidence that the role of vision in the learning processes is finally being recognized as something that greatly surpasses the simplistic organ response to simple light contrast patterns.”(p.577) Furthermore he states “...there must now be the recognition that visually and auditorily directed and monitored movement is the foundation for perception and the skills of discrimination so essential to academic progress.”(p.578) In the time elapsed in which digital

(21)

10

interactive technologies have become the norm, these truths are widely accepted. This has been further evidenced in a wide range of studies, from that of children with learning disabilities (Brill, 2011); to archival work (Mallon, 2015); and to the sequence in which we internalize information and group it into meaningful categories (Otsuka et al., 2013).

As the importance of visualization as a learning tool has become more widely accepted, more attention has been given to the role of aesthetic decisions—such as color selection, symbols, texts, fonts and line density (and many others)—because they all convey visual information that the reader instantaneously interprets on a subconscious level. In geography this is known as qualitative mapping and is a field of growing importance in political science and social studies, because of its ability to show how the spatial subject matter of maps is socially constructed (Leuenberger & Schnell, 2010; Boschmann & Cubbon 2014; Brennan-Horley et al 2010).

Equally important to recognize is that subjective cartographical choices (while potentially rooted in social norms) have practical repercussions to their use—as each choice entails its own fundamental limitation to the way the data observed can be visually depicted. This is the fundamental challenge inherent to all forms of cartographic abstraction and generalization (Isenberg, 2013). Of foremost importance amongst the myriad of visual choices that must be made is appropriate selection of spatial scale, which drastically affects the visual outcome of a map (Schuurman et al, 2007). This decision is such a persistent problem that it has its own term: the modifiable area unit problem (MAUP). Pfeffer et al. (2012: 328) describe this saying “the choice for a particular form, size and location of a geographic area has a considerable effect on the representation of the measured attribute, an effect commonly referred to as the modifiable areal unit problem (MAUP) (Openshaw, 1984b). Several studies have been carried out to get a better understanding of the effects of aggregation and scale associated with the number of output areas (e.g. Holt et al. 1996; Wong et al. 1999).

(22)

11

air of ambiguity hanging over any decision relating to scale, and it falls to the geographer to validate their choice based on the goals of measurement, context of the area, and data source. Scott (2012) is also attune to the importance of scale, which he emphasizes in his economic analysis: “The most obvious and elementary starting point for an investigation of large-scale economic variation across geographic space is to look at the physical landscape and to ask how the opportunities and constraints built into its ecological conditions may have helped to shape a system of human responses.” (2012: 47) Thus at what scale human responses are measured has a tremendous influence on the outcome of the mapping process.

Alongside the selection of scale, an aggregation method and resulting visual output must be chosen that best fit the data set. In modern geographic information system programs, there are a wide range of aggregation techniques to choose from, each with distinct advantages and disadvantages to the data analyzed and the image produced. Mediating the pros and cons of these techniques is therefore very important. Three of the most important established norms in the visualization of data through maps is the use of hot spot mapping for point based data (Chainey et al, 2010); the use of raster grid density maps for continuous data (such as air quality or temperature readings) (Jern, 1985); and using choropleth maps along administrative boundaries for simple demographic representations (Poulsen et al., 2011; Pfeffer et al., 2012).

While there exist longstanding norms for the use of aggregation methods and accompanying visualization, there is a growing debate as to the fundamental importance of the role of space and distance as they pertain to human geography. This debate—previously unthinkable—is a direct result of the effects of digitization, through which the relationship between the individual and space-time have been fundamentally altered. Boschma is one of the most critical voices on this subject in his (2005) work Proximity and Innovation: A Critical Assessment. He argues for a broader interpretation of proximity which includes the cognitive, organizational, social, institutional and geographical forms.

(23)

12

However despite his reservations that “geographical proximity per se is neither a necessary nor a sufficient condition for learning to take place...”(p.62) he must concede that “Geographical proximity is most likely to stimulate social proximity, because short geographical distances favor social interaction and trust building. “(p.67). With this in mind then we next turn to the phenomenon most associated with economic proximity, clustering, whose measurement forms the basis of the analytical portion of this thesis.

3.3 Clustering: Of What, Where, Why and How?

The phenomenon of clustering, because of its inherent emphasis on place, is an important dimension to understanding the dynamics of a city's economic footprint. The primary question that arises is clustering of what? Despite widespread agreement as to the phenomenon of the business cluster, there exists no consensus among academics regarding the benefits or disadvantages to such spatial proximity (Boschma, 2005), types of clusters that exist (Bathelt, 2004; Porter, 2000; Wolfe & Gertler, 2004), nor how to best go about promoting their development (Evans, 2009). Scott points back to capitalism itself as the driving force behind the cluster phenomenon, drawing a parallel between the need of capital to accumulate around nodes of social (and therefore cultural) activity and the tendency of industries to cluster at various spatial scales (Scott, 2012:1). It follows then that cities should be considered superclusters themselves, with layers of interwoven clusters at a myriad of socio-economic levels (Scott, 2012: 174). The inherent complexity of the city can therefore best be understood by looking at the specific pathway of a singular urban space and region (Scott, 2012: 15, 47), and finding a visual method appropriate to capturing its desired condition. Closely intertwined with the study of agglomeration pathways of a specific place, are the types of social clusters that exist and the various

(24)

13

physical spaces that engender them best. Martins' (2015) study of the use of different spaces for varying forms of work activity elucidates just how many facets there are to cluster formation, as secondary and tertiary access points and considerations contribute to the dynamic of a cluster (Figure

1). For example she says, “digital work extends from the office or the residence (the base) to multiple settings (ancillary paces) in what can be defined as an extended workplace” (p.126).

Figure 1: Space Types and their Use for Work. Source: Adapted from Martins, 2015: 130

Scott (2012) also addresses this point with his analysis of spin-off industries which he sees as creating a “...spiral of development” (2012: 23) emanating outward from around a core cluster. Indeed the way in which cluster growth and decline is conceptually visualized is also highly contentious. While Scott suggests a spiral as the metaphoric shape of cluster development, alternative configurations exist. Porter (2000: 20) suggests the diamond (Figure 2) with business context; factor (input) conditions; related and supporting industries; and demand conditions, making up the parameters that predict the likelihood to cluster.

(25)

14

Figure 2: Sources of Locational Competitive Advantage. Source: Adapted from Porter, 2000: 20

Similarly Bathelt and Glückler (2003:132) construct a pyramidal scheme with points of emphasis for economic geographical analysis on the one hand (interaction, organization, evolution and innovation) and a relational perspective (contextuality, path-dependence, and contingency) on the other (Figure 3).

(26)

15

One year later Bathelt's work on the influence of 'local buzz' and 'global pipelines' (2004: 46) constructs an atomic model, with individual actors buzzing around the inside and outside of a core spatial nucleus (Figure 4).

Figure 4: The Structure and Dynamics of Local Buzz and Global Pipelines. Source: Adapted from Bathelt 2004: 46

Most recently Martins (2015: 129) has used a more traditional modulation of a Venn diagram to display the necessary vs. sufficient conditions for a cluster to occur (Figure 5) in a similar vein to Storper's 'holy trinity' whom Batheld & Gluckler (2003: 130) cite as influential (Figure 6).

(27)

16

Figure 5: The Extended Workplace Cluster. Source: Adapted from Martins 2015: 129

Figure 6: Storper's 'holy trinity'. Source: Adapted from citation by Batheld & Gluckler, 2003: 130

Whatever the subject and form of clustering that may occur, or the way its parameters are conceptually defined, the cluster as a concept should be considered an established urban phenomenon, distinct unto itself from any valuation of its measurable effects.

(28)

17

3.4 Global Linkages in Amsterdam and their impact on the CCE

Figure 7: Atlas 1- Flow Lines [Amsterdam: Global Linkages]. (Red= strong) Source: Adapted from Lang et al. 2016

Understanding how clusters interact with their surrounding environment also entails understanding the ways in which cities are linked within larger networks. Increasingly the impact of globalization in everyday life is ubiquitous. In global cities in particular, international supply chains, financial streams, labor pools and social networks stand out as the most dominant forms of interconnectivity today. However there are an infinite number of threads which can be discerned. Visualizations created by Jared Lang et al. (2016) for the Globalization and World Cities Research Network (GaWC) of Loughborough University show that Amsterdam is a city with strong links globally (Figure 7) and within Europe (Figure 8).

(29)

18

Figure 8: Atlas 2- City Views [Amsterdam]: Regional Compacting. Source: Adapted from Lang et al. 2016

However, compared to other global metropoles, Amsterdam falls is in the mid-range. The official GaWC Atlas (Figure 9: Carta, S., & Gonzalez, M., 2010) and their more recent 2012 ranking (parallel to the last year of data collection in this study) lists Amsterdam as a midrange alpha city. Ten years earlier, Taylor and Hoyler (2000)—citing the work of Beaverstock et al. (2000) and their own methodology—rate it only slightly lower as gamma city. Whatever the typology chosen, it is clear that Amsterdam is thoroughly embedded in the global economy through a myriad of strong linkages. Which of these specifically have the greatest influence on the CCE is the next topic of discussion.

(30)

19

Figure 9: α [Alpha] Cities. Source: Adapted from Carta & Gonzalez, 2010.

The CCE is first and foremost an urban economy (Scott 2012: 64). But why is this association so strong? Bathelt et al. (2004: 42) conclude that the impetus underlying global city formation is predominantly social; an urban setting of strong social proximity with a distinctive cultural milieu, simultaneously linked outward through broad macro channels of resource and knowledge exchange. The confluence of these two trends is of defining importance to business formation (especially within a cluster). The authors term the word 'buzz' to define the in-situ benefits that arise when tacit knowledge can only be shared on a face-to-face local basis within a particular social milieu, while 'global pipelines' on the other hand supplement place-specific constraints on knowledge transfer (2004: 32). They argue that these two processes (and the forms of knowledge-transfer they are commonly held to engender) are not mutually exclusive and that the most successful cluster firms are those who make use of both.

(31)

20

Thus they argue for a harmonious balance, in terms of cognitive distance within and between firms, and within and outside the community (2004: 36).

It should come as no great surprise that balance is the remedy prescribed to a business navigating the cognitive-cultural economy—for when is a balanced approach not advantageous? What is more challenging to grasp, though arguably more indicative of the current global structure, is Boschma's argument that spatial proximity is not the foremost condition that fosters clustering. Instead he argues that before the success of a cluster can be attributed to geographical proximity, other forms of proximity must be considered: cognitive, organizational, social, institutional and then geographic (2005: 61). Overall the point of emphasis is once again balance, as firms must navigate the relative pros and cons of each form of proximity emersion. Yet it is the digital revolution—the shrinking of time and distance through the internet—and the resulting internationalization of products and brands that most strongly engender the various forms of proximity. For this reason Boschma argues that the concept of proximity should no longer be constrained by traditional definitions. Therefore it is plausible to argue that contemporarily, geographical proximity might only be a sufficient condition, but not a necessary one, for innovation through other forms of proximity. Therefore neighboring cities may be less connected than distant ones, as indicated by the atlas of Carta & Gonzalez (2010) (Figure 10).

(32)

21

Considering the case of the Amsterdam canal district, it could be theorized that these ulterior forms of proximity are especially important for 1-person firms (which this research shows there are a growing number of) who might otherwise suffer from a stifling amount of both social and intra-industry isolation, as well as vulnerability towards involuntary knowledge spillovers. Such external linkages can enable a businesses to overcome many of the physical constraints of Amsterdam's unique environment.

3.5 Conceptual Model

This research aims to reaffirm the usefulness of data visualization in the form of mapping for economic analysis. An existing gap in academic literature concerning the spatial visualization of the CCE in Amsterdam provides the subject of study. The established phenomenon of clustering (Bathelt et al, 2004; Boschma, 2005; Evans, 2009; Hartley, 2005; Martins, 2015; Porter, 2000; Wolfe & Gertler, 2004) is the conceptual basis which the visualization in this thesis seeks to depict. This research takes a holistic approach (as opposed to an embedded one), because it seeks to contextualize specific geographic and economic sub-cases within the context of the larger case of the city of Amsterdam. This is the subject of the next section, which details the case and data selection techniques.

(33)

22

4) Methodology

4.1 Case Set

This paper began by assessing some of the different academic takes on the still emerging 'new' economy of this century: one defined increasingly by highly-skilled and specialized occupations in hi-tech, business and finance, and creative industries. A growing body of authors contend that global cities are those which rely on this economy as a cornerstone of their urban-economic identity, because it helps to distinguish these cities as attractive destinations for the global highly-educated labor pool (Acuto, 2011; Rekers, 2 012; Sassen, 2016; Scott, 2012). Therefore understanding the location of and composition of this sector within the urban core is important to predicting the future development of the city (Yigitcanlar et al., 2008). At the most macro level then, the case population for this study is secondary Western-European cities of the so-called “third wave” (i.e. post-industrial) (Scott, 2012: 12, 14, 41).

Figure 11 (right): Shape of Connectivity amongst World Cities of

(34)

23

According to the 2010 typology produced by Carta & Gonzalez for the GaWC (Figure 11), there are ten subdivisions of “global” cities in three categories (α: Alpha, β: Beta, and Gamma: ɣ) ranging from largest to smallest in the order of:

α ++, α +, α, α -, β +, β, β -, ɣ +, ɣ, ɣ -.

European cities are well represented in this listing, making for an appealing case-set to select a specific case from.

4.2 Specific Case

Amsterdam was chosen from within this set because it is a typical case (Seawright & Gerring, 2008: 297) of a second tier European α city (figure 11). The premier alpha city of the Netherlands, in Europe, Amsterdam is most comparable to Brussels, Frankfurt, Moscow, Madrid and Milan (Figures 9 & 11, Carta & Gonzalez, 2010). Its population size and bounded physical characteristics preclude it from comparisons with the global α ++ cities of New York and London or the α + cities of Chicago, Paris, Dubai, Hong Kong, Shanghai, Singapore or Sydney. Yet Amsterdam's enduring status as an α city since the 17th century indicates that the city's spatial and structural features—though constraining in some ways—are adaptable to changing social, economic, and technological environments, making it one of the more stable nodes of commerce in Europe. Furthermore the varying roles Amsterdam plays at different spatial scales (regional, national, continental, international) makes the city an appealing case study with linkages that are both pronounced and varied. Goods, people, and information travel from the provincial to the global through Amsterdam creating high-volume node of cognitive, creative, cultural and economic activity.

(35)

24

4.3 Sub-Cases

The Amsterdam Canal District is the predominant sub-case dealt with in this research. The oldest part of the city, its historical pathway and economic footprint are intimately tied to the overall well-being of Amsterdam. When the Grachtengordel deteriorates (as it did during the 70’s) so does the reputation of Amsterdam (Terhorst & van de Ven, 2003: 94). At the finest spatial scale, the Canal District is divided into three distinct parts, West, South and East.

Figure 12: Neighborhood Boundaries of the Canal District. Source: Adapted from Engelen et al., 2013

The city as a whole is divided into eight districts. These are the Center (Centrum) bounded by the Singelgracht (Canal District included); the city’s Western Dock region along the River Ij (Westerpoort); the West, extending from the Singelgracht and Vondelpark westward roughly until Sloterdijk; the New West (Nieuw-West) extending behind the West and below he Western Docks; the South (Zuid) beginning from where the Singelgracht meets De Pijp and the Museumkwartier, and then Southward until the city limit; the East (Oost) containing everything East of the Amstel River; North (Noord) containing everything North of the River Ij; and South-East (Zuid-Oost) a separate administrative area

(36)

25

of the city beneath the east. Practically, the neighborhood and ward classifications of the city provide a more detailed means of breaking down locational activity, by assigning each neighborhood and ward with a specific code. There are a total of 470 neighborhoods in Amsterdam that are derived from within 97 wards.

4.4 Type of Case Study

This research takes the premise that Amsterdam constitutes a typical case within the case-set of European alpha cities, which all constitute the third-wave urban configuration as described by Scott (2012: 64, 67). It is a typical case because Amsterdam has gradually ascended to its status over many centuries and has retained its position through a mixture of local, regional, national and international networks and commitments which keep the city a vibrant hub of economic and social activity in its own right. Because Amsterdam and other smaller Alpha cities are not the major global drivers of the world markets the likes of New York or London, they are more open to penetration by new ideas and economies (Rekers, 2012). Conversely, these cities are also more vulnerable to external economic shocks, which α ++ mega-cities, due to their size, are better able to absorb (Engelen & Grote, 2009). The importance of this contextually underlines why this is an holistic case study.

4.5 Relationship between Cases & Variables

Amsterdam, connected locally, regionally, nationally, and internationally, is economically linked at many different spatial scales, through many different channels (Carta & Gonzalez, 2010; Lang et al., 2016; Taylor et al., 2002). The cognitive-cultural industry economy is multi-faceted and prone to very

(37)

26

different types of input and output conditions. For example banks are not necessarily drawn to a place for the same reason as an emerging artist or tech start-up. Therefore this study aims to chart the individual progression of the three cognitive-cultural branches (hi-tech, business and finance, creative industries) to gain greater insight into their individual composition, any potential overlap between them, and what agglomeration tendencies they exhibit in the city.

4.6 Units of Observation vs. Units of Analysis

The units of analysis in this research are firms. By comparing overall business membership within each CCE sector (business and finance, creative industries, hi-tech) and the extent to which clustering is occurring in similar and different proportions and locations, insight as to the relative economic stature of each sector can be gained. The data used is recorded and stored by the Amsterdam Municipal Office of Research, Information and Statistics (OIS), based on national typologies for business categories as laid out by the Central Bureau of Statistics of the Netherlands (CBS)—therefore it is of rich quality spanning several years which makes it a dependable data-set, suitable for comparison and extrapolation.

The units of observation in this thesis are the different administrative neighborhoods of Amsterdam. With regard to the Canal District, it is distinguished between three constituent ward areas (West, South and East). Meanwhile the neighborhood boundaries of Amsterdam frame the location of the hot spot clusters generated by the ArcMap program.

(38)

27

4.7 Inferences and Predictions

Preliminary inferences derived from the structural layout of the city of Amsterdam and its current economic footprint are as follows:

a) The center of the city inside of the Singelgracht is experiencing an economic boom. Preliminary data used in the early stage of this research also shows that small 1-person firms are growing in number within the city center (Engelen et al., 2013). A casual walk through the center will highlight the lack of empty storefronts and an overabundance of shops selling trinkets and pancakes. One effect of this is that rents in this area have risen dramatically (Tieleman, 2013: 50, 51). This is predicted to have an effect on the formation of clusters within the area, though whether this has increased or decreased their formation is too difficult to predict at the outset.

b) Certain pre-existing neighborhood reputations—such as that of the Old South as a business hub, and that of the Jordaan as cultural area—coincide closely with specific CCE sectors. It is expected that in these cases they will cluster there accordingly.

Overall differences between where each CCE sector locates will reveal the differing needs of each industry, and the parts of the city best suited to support them. At the outset of the research investigation, it is presumed that the three branches of the cognitive-cultural economy will show distinctly different growth and contraction patterns over time, and will generally occupy different parts of the city.

(39)

28

5) Operationalization: Data Selection and

Mapping Methods

5.1 Business Registration Data: Years Covered, Information

Stored, Locational Attributes

The data used is derived from the registration archive of the Amsterdam Municipal Office of Research, Information and Statistic (Onderzoek, Informatie en Statistiek—OIS). However, the definitional categories used to group businesses are decided by the National Bureau for Statistics (Centraal Bureau voor de Statistiek—CBS) and known as SBI codes (Standaard Bedrijfsindeling: Standard Industrial Classification). Within the municipal data set, each business listed is registered with a numeric code according to their professional category as defined by the CBS, along with accompanying information detailing geographic location (either X,Y coordinate, address, postal code, or neighborhood code).

The entire listing is only published every four years. At the time of this thesis, the 2016 data is not yet available, making 2012 the most recent year. Additionally data sets from 2004 and 2008 were used, showing change over an eight year time span. Important to note is that after 2008 the CBS changed the coding numbers and categories (SBI categories and codes) for the data set—in most cases expanding the category ranges within occupations. The effect of this is that business registration codes and definitions from 2012 had to be manually compared and matched back to 2004 and 2008. Fortunately, because this research is interested in three large constructed sub-groupings of the data (business and finance, hi-tech, and creative industries) and the counts of firms within these, changes to sub-categories within those groups do not affect the way the data is processed and visualized. An example of how this works is as follows: in 2012 there are distinct codes for Draadgebonden

(40)

29

Telecommunicatie (cable based-Telecommunications) '6110', Draadloze telecommunicatie (wireless communications) '6120', Telecommunicatie via satelliet (Satellite based-Telecommunications) '6130' and Overige telecommunicatie (other telecommunications) '6190'. However in 2004 all four categories are grouped together under one code for telecommunication '6420'. In both cases the same geographic information is stored, only that it is categorized differently. Because in this research all of those occupations are treated as belonging to the hi-tech sector, their further sub-division does not have an effect on their representation in the data. The matching process described was made significantly easier by the document Schakelschema SBI 1993 – SBI 2008 versie 12-8-08 obtained from the CBS, which shows the full list of the professional categories and corresponding codes for all the data before and after the changes made in 2008.

As a raw data set, there is a variety of information stored concerning firm composition. However there are two issues precluding its full use in this analysis. Firstly information cannot always be transferred through every step of the geocoding process; as points are grouped into polygons, their individual characteristics are sometimes lost. Secondly the information stored—similarly to the SBI codes themselves—is different from year to year. The result of these two factors is that some summary statistics can be provided (such as overall number of firms, 1-person firms, and 'ghost' firms), but only before aggregation. After aggregation only the number of firms contained in a cluster value range can be determined, and within any specific individual cluster. Lastly and importantly, specific firm names and addresses are prohibited from mention as per the confidentiality terms between the University of Amsterdam and the city OIS office.

(41)

30

5.2 Selecting the Amsterdam Cognitive-Cultural Economy from

Government SBI Codes

There are three dependent variables in this research. They are the sectors of business and finance,

creative industries and hi-tech. At times they are abbreviated as BF, CI, and T. These are nominal variable groupings, in that they are not ranked above or beneath each other, but exist equally side-by-side. Importantly, subjective choices were made in determining membership to the variables (sectors), making these categories by no means definitive. Effort was made to rationalize choices based on an interpretation of Scott's definition of the cognitive-cultural economy described in the theoretical framework section (Scott, 2012: 41). What follows are the section headings of the SBI category groups as defined in the English version of the Standaard Bedrijfsindeling (SBI-Standard Industrial Classification) published by the CBS in 2016. From within these groups, specific occupations were chosen. The full listing of specific SBI categories and codes chosen for this research is in the Appendix B, page 108.

Business and Finance

Sections: K—Financial Institutions; L—Renting, Buying and Selling of Real Estate; M— Consultancy, Research and other Specialized Business Services

Hi-Tech

Sections: M—Consultancy, Research and other Specialized Business Services with emphasis on Engineering, Technical Design, Testing, and R & D; C—Manufacturing, with emphasis on electronics (personal and industrial equipment, engineering, repair and maintenance of hi-tech equipment and machinery, support activities in the field of information technology, software production and

(42)

31

publishing, computer management and other ancillary activities). Notable to mention here is the omission of architects from this category, instead grouped within the creative industries, due to the importance of design to their work.

Creative Industries

Sections: N—Renting and leasing of tangible goods and other business support services (emphasis on arts and cultural heritage; tourism, theater performance, production, associated jobs, writing and other creative art, libraries, archives, museums, galleries and exposition centers, preservation work, support funds not in the field of welfare); J—Information and Communication (with emphasis on media and entertainment and publishing); M—Consultancy, research and other specialized business services (with emphasis on creative-services such as public relations, architects, advertising agencies, industrial design, and organization of conferences and fairs).

5.3 Geocoding Data and Creating Choropleth Maps

The data analysis of this research was conducted on the geographic information system (GIS) ArcMap (Version 10.01), published by ESRI. Geocoding techniques are used to link business registration data to spatial areas according to their stored locational attributes. The results are then transformed into different forms of visual representation. The smallest fixed-boundary scale used is a base-layer map of Amsterdam's neighborhoods, of which there are 470 (OIS Amsterdam, 2016). Other than individual house and block numbers, this is the smallest spatial unit the city uses for classification. Each neighborhood is given a two-digit numeric code, followed by a lower case letter (e.g. 33d) which has been made visible on most maps. In some maps (when more practical to describe

(43)

32

location or depict codes with less clutter) the city's ward boundaries are shown instead. These are one scale larger in size, and composites of multiple neighborhoods, whose relationship is represented by shared two digits (ward 33 includes neighborhoods 33a-k).

Aggregating the total number of geographic coordinates of the entire municipal business registration data set within neighborhood boundary lines is the first phase of spatial aggregation. This is called a choropleth mapping, and enables each neighborhood to be given a count value for the number of business entries contained within it. Placing these entries in their proper location within Amsterdam is accomplished through two different geocoding techniques. For the years 2012 and 2008, X,Y, coordinate information exists for each entry. This information can simply be matched to the coordinate grid used to make the Amsterdam neighborhood map. For the year 2004, this X,Y, information was not yet stored, therefore an existing address locator was used for help. This operation uses a type of binary logic through which the ArcMap program systematical deconstructs each address into its constituent parts (house name, number, street name, neighborhood name, and neighborhood code) and then places it in the location with the most matches. The process concludes with the number of exact matches (99% in the case of 2004), along with the number of unmatched entries, and the number of tied entries, which the program could not choose between more than one potential location for the data point. The resulting geocoded information for each of the three years is then color coded according each neighborhood's value, relative to all of the other neighborhoods and their values; with darker colors representing higher counts. From this map the general density of economic activity within a neighborhood can be discerned. Next, the same process is carried out again, this time for each of the three CCE sectors individually in each year. Isolating these sectors from the overall data set is accomplished by selecting only the entries which have a business registration code number (SBI code) that falls within the pre-defined selection criteria of each sectors (Appendix B: 108). Helpfully ArcMap can process such simple coding requests through the “select by attributes” window. This isolates

(44)

33

specific data from the overall set based on a unique attribute, in this case only the specifically desired SBI codes for each sector. As the data has already been geocoded, it can therefore be saved directly to a new shape-file map layer. From this layer, the entries are used to make another choropleth map with neighborhood count values expressed through a color range.

The above described process results in two sets of choropleth maps made: one set for the overall count of all businesses in each neighborhood for the years 2004, 2008 and 2012; one set only for the count of each sector (BF, CI, and T) for each of the three years. Each set is a useful measure in of itself, but the information must be synthesized together to create a meaningful depiction of where the cognitive-cultural sectors are most heavily represented in the city. To determine the relative weight of an economic sector within a specific neighborhood, the total number of specified sector firms there (e.g. T) must be calculated as a percentage of the total number of all firms in the neighborhood. To do this the attribute table of the total data set is joined to each attribute table of the three sectors. In each sectors’ attribute table, a new column is created in which the relative percentage ratio will be calculated. This is a relatively simple calculation (hi-tech count / total count * 100) carried out with the ArcMap field calculator function. Crucially, to correctly perform this calculation, neighborhoods without any data entry points must be omitted, because they are valued as 0 by the field calculator and therefore cannot be divided by themselves as required in the formula. Once this new percent column has been calculated for each sector, it is used as the basis for a new set of choropleth maps. These, like the first set, are aggregated within the lines of Amsterdam's neighborhoods, and color coded with darker values representing higher sectoral concentrations. What is depicted now however is the percent of each cognitive-cultural sector in relation to the full data set of businesses listed in that neighborhood. This is an important step because absolute counts of each sector alone do not highlight whether or not they constitute a small percent or a large percent of a neighborhood's economic footprint. Since the goal of this research is to identify areas of meaningful economic activity where businesses from one or

(45)

34

more CCE sectors reside, first the areas where these sectors make up a significant statistical percent of total economic activity must be identified.

5.4 Isolating Neighborhoods with the Highest CCE Activity

While calculating the percentage-ratio of each CCE sector within each Amsterdam neighborhood is useful to understanding where in the city meaningful agglomerations of hi-tech, business and finance, and creative-industries are located, it is not enough to rely upon them for the cumulative analysis of this work. This is because percentage ratios can also be deceptive in information they convey. This occurs because percentage ratio is not an indicator of actual count size. For example, a neighborhood might have a composition ratio of 100 percent for the hi-tech sector. However this metric does not necessarily mean that this neighborhood is a meaningful center to hi-tech activity. There might only be one registered firm in total for that neighborhood. If that is a hi-tech firm, than it is assigned a value of 100 percent. However in reality this neighborhood is a statistical outlier. This problem becomes acute when a color scale is used to define the percent ratio, which leads to such outliers being misrepresented and grouped alongside much higher-count areas of activity. To solve this problem a visualization is needed that shows both percentage ratio and the respective count of a CCE sector for each neighborhood.

A two-step approach was taken to address this problem. Firstly the neighborhoods contained within the highest ratio-percentage range for a specified CCE sector are isolated and saved as a new distinct map layer. However, with the exception of the business and finance sector, this top percent range made up too small a portion of the data sets of hi-tech and creative industries. Therefore, for those two sectors, the second-highest percent range was also isolated. These new layers, also in choropleth form, are given distinct colors to differentiate them from one another when they are viewed

Referenties

GERELATEERDE DOCUMENTEN

Als namelijk alleen de grootte verandert, kan men, mits de verhouding waarin de buffers belast worden bekend is, ervoor zorgen dat elke buffer evenveel

Purpose: To determine if prostatitis and prostate cancer (PCa) can be distinguished by using apparent diffusion coefficients (ADCs) on magnetic resonance (MR) images, with spec-

I found that high audit tenure, associated with high audit fees, has a significant negative effect on the value of discretionary accruals, which means that the

To summarize the main theories of CSR and their influence on sales: CSR influences the sales of products through different strategies. The main strategy for sustainable

Analogously, the cultural system (note: not “a culture” yet, we will attend to this below), processes actions as communication leading to changes in

This either means that these topics, in collaboration with the cultural city regions, have been moved in responsibility from the national to the regional level or that the

In the coalescence channel, the time for adsorption to take place (droplet formation time and adsorption time in the meandering channel) is relatively long and can be used to

LIST OF FIGURES Figure 1.1 Administrative Hydrological Regions Figure 1.2 Structure and constituencies of Mexico’s river basin councils Figure 1.3 River Basin Councils in Mexico