A central question to the MAKE-IT project is the Maker movements’ near and long-term potential to actually disrupt current production and consumption patterns. Of course, that is a multidimensional question whose complexity can’t be comprehensively reflected in a single sentence. Disruption is better understood as a process rather than an event, if it is to happen at a societal level. The process of disruption starts with becoming aware of a situation which is not satisfying anymore. However, this realisation will depend on persons’ position within the economical ecosystem. Incumbents of lucrative positions within the current system will likely oppose emerging solutions pointing at their shortcomings, often related to a lack of mature and robust implementations of otherwise progressive, forward-looking ideas such as fabricating and assembling products on-site and thereby avoiding shipping and related burdens for the environment. Similarly, there is an argument that making enables people to become more entrepreneurial and therefore create their own jobs, quasi employing themselves.
In both situations, we rely – among other things – on the idea of networks as empowering distribution mechanisms of either product knowledge (e.g. 3D models, materials and related parameters for desktop production) or ‘how-to’ knowledge (e.g. instructions, shared experiences and evaluations of alternative solutions). Hence, one of the ‘signs of disruptions’ is the emergence of large-scale networks empowering makers to share and comment their knowledge. In this posting, we want to more closely study the relationship between network dynamics, as reflected by their increasing membership and related network characteristics such as various centrality measures, the degree of members (often referred to as ‘connectivity’) or the largest sub-networks within a community.
The role of online platforms
By analysing existing online platforms such as thingyverse.com or instructables.com, we aim to contribute to a better understanding of collaboration and connectedness within the Maker community. Eventually, this will be of help when pursuing the realization of the movement’s real potential due to the identification of missing features or incentives that stop platforms from supporting desirable behaviours. Or, alternatively, we might realize that the existing platforms serve a distinctly different purpose and are therefore less suitable to support the larger picture of digital, social innovation as envisioned within large parts of the Maker movement.
However, a note of caution is in order, since we do not claim that online platforms, though arguably an important element, are representative of the entirety of the Maker movements where still a lot of interaction happens offline in places subscribing to different formats (Makerspaces, FabLabs, techshops, hacker labs, wet labs etc.). Still, online platforms help gathering like-minded people, sharing experiences and exchanging ideas. For the moment, we look at each platform separately since they offer unique sets of functions and collect different data, which makes it tricky to apply questions across the board and aggregate insights obtained from diverse platforms.
The case of instructables.com
Instructables is a website specializing in user-created and uploaded do-it-yourself projects, which other users can comment on and rate for quality. It was launched in August 2005.
In this blogpost we concentrate on first experiences with instructables.com. In 2011 the platform was obtained by Autodesk, a corporation that produces software for the architecture, engineering and construction business and is best known for its software AutoCAD. Projects (also referred to as instructables) are mainly published as step-by-step instructions, comprising text and images. Data on authors and their projects can be accessed via an API. We obtained a dataset of 225.681 projects published by 74.824 distinct first-authors. Each project is organized into one out of eight categories, with technology, workshop, and craft being the most popular categories (see Figure 1). The number-ratio of published projects in any two categories remains roughly constant over time, where the category play relatively lost importance in the last few years and the publication rate in the category costumes has a pronounced seasonal dependence. At a certain time each year, the number of projects published in the category costumes are in vicinity of, or even outnumber the projects published in the category technology. At a more detailed level projects are organized into channels and can be described with key-words. Arduino is both the most frequently used channel and the most frequently issued keyword, which again highlights the communities technical affinity.
The long-term activity clearly indicates growth, but also shows signs of strong fluctuations. For example, we cannot provide explanations for the drop of uploads at the end of 2014 / beginning of 2015 or the decline in the first half of 2017. Overall the movement is still growing, while the relative distribution of categories did not change with time.
Only 0.68% of all projects are created collaboratively
Returning to our initial question of how much collaboration or interaction can be observed within Maker networks, in this case the network of inctructables.com members, we were looking for behaviours such as liking, favouring or downloading each other’s’ projects. Platforms such as thingyverse.com go even further and collect information of how many members actually ‘made’ or implemented an object. A prove of the replicability of the shared knowledge.
Surprisingly, despite the apparent openness and inclusiveness of the community, collaborative co-authoring of projects is not popular. Based on our sample, only 0.68% of all projects are created in collaboration, this suggests that most makers operate on an individual basis and do not form large collaborative networks to perform projects. These collaborations are functionally supported yet not incentivized by instructables.com, as co-authors are not given any credits on the web-frontend.
The small amount of collaborative efforts can be studied in terms of networks, where users are represented as nodes that are connected by edges as soon as they publish a co-authored project. As only a small minority of the users co-authored a project, only few nodes are connected with edges and not every user can be reached through a chain of contacts from other users. This situation is similar to performing the small world experiment, where most people don‘t know anyone – messages won‘t come across the network. The opposite is found in an experiment by Milgram, involving a much larger community, who probed the personal acquaintance-network of people living in the USA and found the well-known result of 6-degrees of separation between any two people. Parts of a network, where each node can be reached from any other node via edges are called connected components.
Analysing the increase of connections within a large sub-network
We created a video of the largest component (comprising the most nodes) to learn more about the dynamics in co-authorship networks. The date is in the top right corner. A node is added to the network after a user published his/her first project. Nodes are rendered red, if the respective user published a project at the given point in time. The color changes subsequently to light-red (last publication within the last month), solid grey (last publication within the last four months) and light grey, the edges are rendered red at the point of publication and light grey thereafter. The video shows a naturally grown structure of locally organizing makers.
A note on licensing:
While licensing represents a possible caveat, as this invites or denies for reuse and contribution, it does not seem to have a big impact on any of the collaboration processes. The majority of all projects (84.6%) are licensed under the Creative Commons Attribution – NonCommercial -ShareAlike license, only 4.4% of projects had ‘All Rights Reserved’.