2.1 Models of maker roles and identities

The literature review of this section revolves around conceptual models of the Maker Movement presenting conceptual maps as another type of exploratory maps. These are models of who and where makers and their laboratories are and how they could be found, analyzed and understood, in a sense-making effort for orienting their exploration. The first model depicts the roles and identities of makers as the starting point for understanding the nature, distribution and social dimension of the Maker Movement.

The definition of maker proposed by Dougherty and MAKE Magazine is broad and fuzzy enough to promote the growth of a global phenomenon, but less clear for analyzing and organizing it.³ Chris Anderson improved the definition with specific practices and principles based on: a) digital design and prototyping; 2) a shared practice of collaboration and sharing of projects; 3) digital fabrication technologies and spaces.⁴ Makers can be considered (and often are) designers or a new kind of designers, be them formally trained and employed or informally active and self-taught. Often working with open, peer-to-peer, distributed and Do It Yourself (DIY) approaches in a collaborative way,⁵ makers adopt digital fabrication technologies⁶ and work often for cultural change,⁷ educational⁸ and social⁹ purposes, beside entrepreneurial ones.¹⁰

If anybody can be a maker,¹¹ then the identity of a maker is likely to result from the integration of different profiles, roles, knowledge, practices and identities. Therefore, an exploratory approach to start addressing the complexity of makers’ profiles can be developed through a simple model¹² that enables to view the roles and identities of makers in a more nuanced and layered way. Within this model, the makers’ identity is the result of the integration of different roles and practices that concur for the same common purpose of their initiatives (fig. 1). Depending on the nature of the common purpose, several versions of this model of makers’ identity could be elaborated: for example one for makers working with social innovation; another one for makers working with commercial purposes or for cultural purposes, and so on.

2.2 Places, communities and scales of the Maker Movement

A second map for navigating the Maker Movement can be drawn for representing a model of the different communities that can be found in the movement, and how they interact among each other and at which scale they operate. Such communities can be found on three levels and with a cross-cutting socio-technical dimension of digital platforms (fig. 2):

1. Local: communities that form in and around local laboratories and events.

2. Global: a global community emerging all the local events and laboratories. This will be explored in depth in section 4, especially with an updated analysis (4.2).

3. Projects: the communities that form around the development of projects which are typically prototyped and manufactured locally in the laboratories; projects could also be completely global, especially when developed digitally in a common repository.

4. Digital platforms: a cross-cutting dimension that connect the previous three scales, for example for sharing projects openly as Open Design, which then become community-based initiatives. This will be explored in section 4, in the literature review (4.1).

2.3 ​Models of Maker Laboratories

Different practices and roles constitute who makers are and how their communities work at different scales and dimensions: as a result, we find different places with specific traditions, practices, technologies, business models and values (fig. 3).

Roots of Maker Laboratories can be tracked back to the Hacklabs movement, which started around 1995 and peaked around 2005: voluntary-run spaces that provided access to machines with GNU/Linux and Free Software, programming and electronics as a physical and local platform for developing and promoting the hacker culture, autonomous movement and media activism, post-capitalism and alter-globalization politics. Hacklabs typically have a strong political agenda, and they share overlapping but different trajectories with Hackerspaces, less politically inclined and more widespread.¹³ Hackerspaces started as well in the 1990s, and went through three waves, and only recently they flourished with a stronger effort on the organizational dimension and focus on the physical dimension of hacking, from microelectronics to 3D printers.

Fab Labs emerged, almost by accident, as the outreach and democratization of the work of Neil Gershenfeld at MIT on digital fabrication as the key translator between bits and atoms, of information and matter.¹⁴ This is the main focus of Fab Labs, at least at the technological level: it is not by accident that the concept has been developed at MIT at the Center for Bits and Atoms,¹⁵ not at the High-Low Tech¹⁶ research group. Fab Labs offer a partially defined and shared set of tools, processes and knowledge for developing physical representations of digital data, for extracting digital data from physical contexts and for experimenting how digital technologies can influence the development of physical objects. Furthermore, the Fab Lab community focuses on sharing same protocols, practices, communication channels and initiatives in order to enable any project to be replicated everywhere.

Makerspaces are similar places, but with less focus on the digital dimension of making and more on the analog one; they are sometimes considered as the main term for representing the whole formats of Maker Laboratories. However, sometimes they can be clearly defined as a separate community from Fab Labs and other labs, and are often promoted by MAKE Magazine.¹⁷ Several different approaches at organizing Makerspaces can be found: one example is the now-defunct network of TechShops, who were typically ten times larger than a Fab Lab and with a more entrepreneurial approach than a community-driven one.¹⁸ Sewing Cafes¹⁹ are similar places but with the goal of enabling their members to work with textiles and fashion. Repair Cafes, emerged in Amsterdam in 2009, are also related, but with a specific focus not on making but on fixing existing commercial products for local neighbors.²⁰

The same hacker ethic of sharing knowledge, free access and collaboration²¹ has been adopted and promoted by the DIYBio Movement and its members:²² DIYbiologists have expertises and experiences different from the makers’ ones, but share with them some common principles, representing thus a similar culture and community. The DIYBio Movement aims at democratizing access to research in biotechnology, and this approach is increasingly adopted within the Maker Movement, for example for the design of 3D printers that employ orange juice and modified bacteria instead of plastic filament.²³ This connection between the Maker Movement and the DIYBio movement has also been fostered by MAKE Magazine and Maker Faires with the perspective that biology could be considered as a 'personal technology' just like making and digital fabrication.²⁴

A third conceptual map can be drawn to communicate the dimension and connections among such type of laboratories. This is a preliminary, hypothetical map based on the experience of the authors, who have not only been just researchers and professionals of the Maker Movement, but also participants of several of its communities for years (fig. 3). Fab Labs, Hackerspaces and Makerspaces are shown with a larger size since they are the main formats in terms of status of development, distribution, number of places and popularity. Because of this, they often tend to include other spaces (Repair Cafes, Sewing Cafes, DIYBio Labs) or at least part of their technologies, practices and communities. The overlaps between these formats are related to the making and entrepreneurship activities that connect them, building opportunities for recognition, the construction of communities, the exploitation of value created, as well as possibilities for replication of social innovation initiatives. As a consequence, the emerging landscape is a holistic ecosystem of value creation for societal change. Section 4.2 provides a data analysis for the validation of such simple and conceptual map.

3.1 ​Literature review

Maker Laboratories often act as (or are part of) schools, community hubs and professional centers where the Maker Movement has been emerging and building social and collaborative initiatives. Because of the often bottom-up nature of the movement, the number of these laboratories is always changing, making it difficult to be completely tracked over time. Their amount is of strategic importance as it could be considered as a proxy of the overall number of makers by considering each local community, through the estimation of average quantities or by directly contacting laboratories to receive a more accurate estimate.

The “Makers’ Inquiry” initiative²⁵ aimed to explore the emergence of the phenomenon in Italy, in its first years. It proposed that the geographical distribution of the Maker Movement could be assessed from different places where makers “make” their making activity (home, office, co-working, workshop, artisan workshop, school or university, factory, Maker Laboratories and so on). The most interesting fact is that these activities are carried out in a range of different places and that these places could be complementary to each other: not only Fab Labs, but also schools, studios and at home. Most of those laboratories were found to be located in North and Central Italy, and therefore may be directly linked to the local industrial traditions, and they are hosted in places more related to crafts, business and production rather than research and education. A similar distribution was found in another research about Maker Laboratories in Italy.²⁶

Another initiative analyzed laboratories in France,²⁷ exploring it with a "tour of the labs" experience. It is not only as a way to identify the distribution of Maker Laboratories, but also to get in touch with makers’ peers, learn new practices, while helping to animate the network.

The Joint Research Centre (JRC), the European Commission’s science and knowledge service, recently elaborated a report that proposes an overview of Maker Laboratories at European level using several sources of data. According to this analysis, Fab Labs account for nearly half of the laboratories in the European Union (48%; 397 laboratories), Hackerspaces are 40% (327 laboratories) and then there are other types of laboratories for 12% (102 laboratories). The average number of laboratories per country is 29.5. France, Germany and Italy represent 53% of the laboratories within the European Union.²⁸

An analysis of the hackerspaces.org platform found that Hackerspaces are a global phenomenon in 71 countries but with a greater presence in Europe and the USA.²⁹ The majority of labs are in the USA with 238 labs, with Germany at the second place with 131 labs, then the United Kingdom (51), France (42) and the Netherlands (28). Brazil has the largest number of Hackerspaces in South America with 28 labs, and China the largest in Asia with 26. South Africa is the largest in the African continent with 4 labs.

3.2 Data analysis: the geographical distribution of Maker Laboratories on fablabs.io, hackerspaces.org, diybio.org

In order to advance the mapping of the geographical distribution of Maker Laboratories, in this section we present the analysis of an already formatted and openly accessible dataset that collected data from the fablabs.io,³⁰ hackerspaces.org,³¹ diybio.org,³² platforms on January 25, 2018. Such dataset was created with a custom software module that accesses and standardizes data from many Maker platforms in order to produce a common set of APIs.³³

In this section we plot the geographical distribution of these labs at global, continent, country and major city level (fig. 4-7). At global level (fig. 4), the majority of the laboratories are Hackerspaces (2,237 labs), almost the double of Fab Labs (1,216); DIYBio Labs (104) are a minority. This could be both a measurement of their popularity, but also of the efficiency of the platforms in mapping them, or the quality of the gathered data (hackerspaces.org is a wiki and can be edited by anybody, the other two platforms have an editorial team).

At continent level (fig. 5), North America and Europe are the main places where labs can be found, but while in North America Fab Labs are one fifth of Hackerspaces (220 labs and 850 labs, respectively), in Europe they reach similar numbers (630 and 696 labs). In Asia Fab Labs are the majority (216 and 90 labs) as in Africa (50 and 47), and in South America again Hackerspaces are the majority (118) but Fab Labs are very close (91). The highest concentration of DIYBio Labs is in North America (48) and Europe (38).

At country level (fig. 6), USA has the main concentration of laboratories (352 Hackerspaces, 170 Fab Labs, 38 DIYBio Labs); China has the second place for Hackerspaces (212), has 21 Fab Labs and there is only one DIYBio Lab. Germany comes third for Hackerspaces (185), but has only 47 Fab Labs and 5 DIYBio Labs. It is interesting to note how we can find a majority of Fab Labs instead of Hackerspaces in France (155 and 76 labs) and Italy (134 and 33 labs).

At city level (fig. 7), the majority of cities have more than three laboratories; Europe also concentrates a huge amount of laboratories, but there are interesting exceptions, like Tbilisi with 11 Fab Labs at the first place. Main cities can be considered Paris (9 Hackerspaces and 8 Fab Labs) and Shenzhen (7 Hackerspaces and 5 Fab Labs). New York and Los Angeles have the same number of Hackerspaces (9), but no other labs. In terms of Fab Labs, we should note Boston (8), then Lima and Milan (6), then Sao Paulo, Shenzhen, Quito, Porto Alegre, Madrid, Dubai (5).

Generally, the Global North concentrates more laboratories than the Global South. The areas with the highest concentrations of laboratories are Europe, the East Coast and Midwest in the USA, South of India and Brazil, whereas China has Hackerspaces distributed all over the country.

4.1 Literature review

Digital platforms, analyzed with a social network analysis, are the main source for all the overviews of the social structure of the global community of Maker Laboratories and in-depth analysis of specific maker projects elaborated so far. The ubiquity, scale, ease of use, quantity and quality of available data render platforms not just important for users, but also for researchers.

For example, an analysis of the global community of Hackerspaces considered hackerspaces.org as the main source for the data, but did not explore the geographical distribution of labs but their social structure.³⁴ The list of labs presented on the platform (a wiki) was analyzed in their connections by in-degree, first at depth 1 (considering only links between labs listed on the platform) and then by connecting the labs found at depth 1 with the ones originally listed in the platform but not directly included or connected to it. In the first case, 941 labs were found, structured with a densely connected centre surrounded by a concentric distribution of less connected labs: two main labs can be found at the core, ccc.de (Chaos Computer Club), located in Germany and the noisebridge.net (the Noise Bridge) located in San Francisco in the USA, with Metalab (Austria) between the two. These labs are surrounded by the ones from their same country first and then by the labs from other countries, belonging to Europe, The United States and Canada, South America, Australia, Asia and Africa. In the second case, 1,034 labs were found, with a more clustered network with more links at the country level, and a strong sub-community from Germany made of Fab Labs and not Hackerspaces. Both analyses found also the presence of several Makerspaces, showing thus how the boundaries between these formats of labs are not so clear, even on platforms with a clear identity such as hackerspaces.org.

Another analysis focused on the whole global community of Maker Laboratories (Fab Labs, Makerspaces, Hackerspaces) by using instead Twitter as the main source of data, mined with a custom software released as open source that analyzes who follows whom in a manually-curated list of Twitter accounts of Maker Laboratories.³⁵ Here Twitter is considered as a proxy of the global connections among the laboratories, and the obtained network consists of 946 nodes and 29,821 edges among them, representing thus a first data-driven measurement of the size of the global community. The community is split into two main polarities, and an exploration of its sub-communities at a high level of resolution³⁶ shows that Hackerspaces, Makerspaces and TechShops are grouped together on one side (53.28% of the nodes), and Fab Labs on another side (42.07% of the nodes), with a subset of French Fab Labs as a separate sub-community, showing a first subdivision of the Fab Lab community. At a finer resolution, the number of sub-communities rises, clearly showing the distinctions between Hackerspaces (32.66 %), Makerspaces (16.7 %) and TechShops (1.48 %). Within Fab Labs, however, many more sub-communities can be found, suggesting that form a much more diversified and articulated network of laboratories. Generally, few nodes have high degree and betweenness with a common power-law distribution, and this can be found also regarding influence and trust, which have a different meaning in each polarity: Eigenvector centrality is much more concentrated in several nodes in the Fab Lab community, and PageRank centrality is concentrated in very few nodes among Makerspaces and Hackerspaces. A static perspective on trust finds it to be distributed among many labs in Fab Labs; a dynamic perspective finds it concentrated in very few nodes on the Makerspaces and Hackerspaces side.

Within the Maker Movement, projects can be at both local scale and global scale: leveraging the principles and the attitude of Open Source Software, projects can potentially scale up to many participants, but typically they are small projects that start from the local context, since the hardware dimension renders upscaling much more costly and cumbersome. As in many research initiatives about Open Source Software, projects are analyzed through their hosting on version control systems like Git and their platforms like GitHub. A large scale social network analysis of Open Source Hardware projects (105 projects) was developed in order to understand to which extent the transparent and participatory processes of software development reached hardware product development: the result is that these initiatives are generally small-scale and heterogeneous.³⁷ Social network analysis has also been adopted by makers/researchers in order to understand their participation in open and maker projects, and their position in the networks of interactions emerging from the collaboration in GitHub in defining Open Design, teaching it and developing a maker platform for Open Design projects. These are all meta-design activities that build a socio-technical infrastructure of Open Design projects, rather than directly designing Open Design projects.³⁸

4.2 Data analysis: An update of the global structure of the Maker Movement on Twitter

The software developed for the analysis of the whole global community of Maker Laboratories on Twitter mentioned in the section above³⁹ is openly accessible, and therefore we adopted and updated it in order to replicate the same analysis five years later and compare how the community has evolved so far. Here again the accounts were manually added in another updated list and analyzed in terms of who follows whom, as a proxy for collaboration and trust among the Maker Laboratories, resulting in a larger network of 1,278 nodes and 52,533 edges.

Sub-communities can be observed with the same algorithm⁴⁰ adopted by the previous research, providing different resolutions that enable the uncovering of network structures at different scales. With a resolution of 1.0, we can see that the larger part is made of Hackerspaces and Makerspaces (45.07%, blue nodes on the right), followed by Fab Labs (31.61%, red nodes on the left), then by French Fab Labs (10.02%, orange nodes on the left), then by Maker Faires (7.36%, light blue nodes on the right) which are closer to Makerspaces and Hackerspaces than to Fab Labs (this might be a stronger connection of MAKE Magazine to Makerspaces than to Fab Labs). It should be noted how Repair Cafes, at 1.49%, are a separate branch on the top left, and that there is handful of completely disconnected labs, mainly Makerspaces and Hackerspaces (fig. 8).

At a smaller scale and finer network architecture, with a resolution of 0.5, more sub-communities can be found: Hackerspaces (24.41%, orange nodes on the right) and Makerspaces (15.26%, yellow nodes on the right) are again separated at this level, but now they are not so defined, as labs can be found in either part. We can then observe French Fab Labs (9.7%, pink nodes on the middle left), Italian ones (6.96%, light blue nodes on the left), followed by Maker Faires (5.87%, yellow nodes on the right) and only later by Fab Labs from mixed countries (5.71%, dark blue nodes on the bottom left) (fig. 9). TechShops are here now part of the Makerspace community, and while the French Fab Lab community was almost separate already five years ago, the Italian community has now emerged as a more identifiable entity now.

The distribution of Degree centrality (i.e. the number of edges of a node—the more the edges, the higher the centrality in the network) shows a larger concentration of high degree centrality in the Fab Lab community (fig. 10).

The distribution of Betweenness centrality shows very similar results compared to the previous study, pointing out how still very few nodes bridge the two polarities (fig. 11). Betweenness centrality measures how many times a node acts as a bridge along the shortest path between two other nodes, i.e. how many nodes it can bridge.

The distribution of Closeness centrality is rather homogeneous in the network, with rather high values shared and very few nodes with a very high value. Closeness centrality measures the distance (shortest paths) between a node and all other nodes in the network, i.e. the closer a node is to all other nodes, the more central it is. Almost all nodes are therefore very close to each other and able to spread information efficiently (fig. 12).

The distribution of Eigenvector centrality shows how trust and influence is more concentrated in the Fab Lab community, as it was previously found, but also rather fairly distributed among most of the labs (fig. 13). In Eigenvector centrality a node is important (central) if it is connected to other important nodes.

As for Betweenness and Eigenvector, the distribution of PageRank centrality shows very similar results to the previous study: trust in the network has stayed almost the same (fig. 14). PageRank centrality is a variant of Eigenvector centrality: here influence is determined with an iterative approach where nodes vote for the importance for other nodes (influence is calculated with iterations of voting over connections instead of connections only).

Overall, the network has become slightly larger, with Maker Faires and Repair Cafes emerging, and TechShops disappearing, and the Fab Lab community split between French, Italian and other countries (and the last two groups are more integrated into each other than the French part). The distributions of the centrality measurements have remained similar to the previous analysis.