First Monday

Digital nomads, coworking, and other expressions of mobile work on Twitter by Jeff Hemsley, Ingrid Erickson, Mohammad Hossein Jarrahi, and Amir Karami

We analyze a set of Twitter hashtags to ascertain how contemporary parlance in social media can illuminate the rich cultural intersections between modern forms of work, use of technology, and physical mobility. We use network word co-occurrence analysis and topic modeling to reveal several thematic areas of discourse present in Twitter, each with its own affiliated terms and distinctive emphases. The first theme centers on worker identity and is currently dominated by the experiences of digital nomads. The second theme focuses on the practicalities of working in a physical location and is currently dominated by issues related to co-working spaces. Finally, the third theme is a loose and speculative set of ideas around the evolution of work in the future, predicting how enterprises may have to adapt to new ways of working. Our contribution is twofold. First, we contribute to scholarship on social media methods by showing how a robust analysis of Twitter data can help scholars find subthematic nuance within a complex discussion space by identifying the existence and boundaries of topical sub-themes. Second, we contribute to scholarship on the future of work by providing empirical evidence for the ways that the myriad terms related to mobility and work relate to one another and, most importantly, how these relations signal semantic centrality among those who share their thoughts on these types of work.


1. Introduction
2. Literature
3. Methods
4. Quantitative results
5. Qualitative results and discussion
6. Conclusion



1. Introduction

The shifting nature of knowledge work has been a ripe area of study for decades, closely mirroring the tight coupling between advancements in technology and new ways of working (Orlikowski and Scott, 2016; Ens, et al., 2018). These technological advancements, particularly with the introduction of digitized management information systems, ushered in the notion of ‘informating’ work (Zuboff, 1988). More recently, the Internet has enabled knowledge work to divorce itself from fixed locations (e.g., hardwired desktop systems) and, in turn, has created new arrangements of networked collaboration (Schmidt and Bannon, 1992). The introduction and growth of mobile technologies as an integral part of knowledge work is witness to the continuing development of these new arrangements (Ojala and Pyöriä, 2018; Francisco and Klein, 2018; Erickson and Jarrahi, 2016). The resulting mobility of knowledge work is a concatenation of all that came before (de Carvalho, et al., 2011; Sørensen, 2011). It is digital, networked and typically collaborative, in addition to being decoupled from a specific work location.

As a result of this history, researchers interested in studying this phenomenon begin by facing a cacophony of terms in the literature (e.g., mobile, nomadic, location-independent or even remote work). They quickly see that there is little linguistic consensus related to the organization and experience of working away from a traditional office space (Costas, 2013; Sørensen, 2011). This conceptual ambiguity, if problematic, is to be expected as a sociotechnical phenomenon such a mobile work emerges and matures. However, what we find more troubling as scholars in this area is that within this period of sensemaking, little to no attention has been paid to how workers themselves describe their own experiences of mobility. This paper seeks to address the entwined topics of mobility and work from a “user” perspective — that is, by looking to how workers, those that support workers (i.e., managers, service providers, etc.), and other interested players (i.e., members of the media) naturalistically discuss these topics on social media. We believe that this perspective not only adds complementary value to extant academic scholarship but may also assist researchers in the field with new empirics that can help to differentiate certain sub-themes within the larger discussion from one another.

Social media sites like Twitter have been recognized as places where influential actors, through their discussions and online posts, play an important role in setting cultural agendas and norms (Chadwick, 2013; Neuman, et al., 2014). They often use hashtags to establish or signal an affiliation with a larger social conversation (Bruns and Burgess, 2011; Small, 2011) or provide topical context (Golder and Huberman, 2006; Marwick and boyd, 2011). As such, analyzing hashtags and how they are used individually and in groups reveals an important dynamic regarding the way that cultural meaning is shaped (and reshaped) within social media (Blaszka, et al., 2012; Cunha, et al., 2011; Ma, et al., 2012; Bozzi, in press).

Using network word co-occurrence analysis in conjunction with topic modeling, we discover that there are several thematic areas of discourse around mobile work present in Twitter, each with its own affiliated terms and distinctive emphases. These linguistic divisions underscore how Twitter users appear to be reinventing, repurposing, or rearranging existing terms to describe or comment on shifting work arrangements. Specifically, we find that there are three distinct themes. The first centers around the hashtag digitalnomad, and the surrounding conversation tends to be related to workers who constantly travel and stay digitally connected in order to work productively. The second thematic space centers on the tag coworking and tends to include posts related to working in shared, often rentable, work spaces. The final theme revolves around the tag futureofwork and includes in its orbit tags like hr, flexibleworking and remotework; it appears to draw its coherence from discussions about enterprise organizations and their need to adapt to emerging work trends.

This work makes a contribution, first, by showing how a larger discussion space on Twitter might be analyzed to identify and surface a set of embedded sub-themes. Second, we provide new insights about the semantic centrality of several terms related to mobile work from the prospective of those who are engaged in this work.



2. Literature

Generally speaking, the phenomenon of ‘mobile work’ comprises the movements of knowledge workers, their chosen locations of work, the collaboration practices accomplished on the move, and the embedded networked infrastructures that enable this professional form of mobility (Koroma, et al., 2014; Sørensen, 2011; Cousins and Robey, 2015; Sutherland and Jarrahi, 2017). Scholars from information science, management science, and sociology, among other disciplines, have all looked at precursors and close equivalents of mobile work (e.g., distributed work, remote work, telecommuting, crowd work) in prior research, but so far have refrained from defining the phenomenon as wholly distinct from other forms of work. This may be, in part, a function of nomenclature; several different words are commonly used to refer to aspects of mobile work. Mobile workers are sometimes referred to as being ‘nomadic’ (de Carvalho, et al., 2011; Harmer and Pauleen, 2012; Pinatti de Carvalho, et al., 2017; Su and Mark, 2008; Thomson and Jarrahi, 2015), a description that stresses frequent, if not constant, movement. The key emphasis here is on a worker’s locomotion; he or she is imagined as always being in motion, untethered and unmoored. By contrast, work and workers are other times labelled as being ‘remote’ (Bailey, et al., 2015; Dix and Beale, 1996; Koehne, et al., 2012; Rockmann and Pratt, 2015), which is a designation that zeroes in on location, typically fixed. To be remote is to be away — usually pointing to an alternate work location, such as a home, satellite office, coffee shop, or coworking location, as the most salient identifier of the worker (Spinuzzi, et al., 2019; Leclercq-Vandelannoitte and Isaac, 2016; Thompson, 2019). Remote workers, by this definition, are elsewhere, connected back to base (often a central organization) by virtue of networked technologies (Wood, et al., 2018; Charalampous, et al., 2019).

Recently, a few new words have crept into the lexicon relating to mobile forms of work. For example, the new phrase work in place (Erard, 2016), while emphasizing location, de-emphasizes its relation to any separate central setting. Key to this label is a plea for all work locations to have equal legitimation; the term ‘remote’ tends to reinforce a hierarchy of work locations and, more often than not, to the primacy of work that occurs in those locations. Workers who ‘work in place’ are not distant others, one editorial writer claims (Erard, 2016), but in the right place for their own (and presumably their company’s) productive aims. In other words, all places can be workplaces and all workplaces should be considered equal.

Another new term, digital nomad, is also on the rise (Hart, 2015; Johanson, 2014; Spinks, 2015; Gretzel and Hardy, 2019; Reichenberger, 2018; Lee, et al., 2019). This label is an extension of the earlier term ‘nomad’, and has quickly come to cultural recognition as a referent for an emerging class of workers whose members maintain their employment while being both constantly in motion and highly remote (Orel, 2019). As commonly conjured in popular media, most digital nomads are millennials intent on blurring work/life boundaries by leveraging technology to position themselves on both fronts as advantageously as possible (Thompson, 2019). In other words, they appear to be working while perpetually traveling (Schlegelmilch and Lysova, 2018; Nash, et al., 2018).

One of the challenges for scholars working in this conceptual area is a lack of clear definitions for these terms (Ciolfi and de Carvalho, 2014). Often there is an assumption that past studies sufficiently describe the nature of and language surrounding mobile work. Our research approach does not make this assumption, but rather seeks to disentangle the many existing terms to understand their semantic boundaries and relations. We take the view that the phrases people use in making public broadcasts on social media reflect broader cultural conversations.

This is not a new idea. Language has been seen by linguists, in particular, as highly representative of culture for millennia (Carey, 2008; Gee, 2015). Use of metaphor, for example, evokes vivid imaginings in people that go on to structure subsequent expectations or interactions (Lakoff and Johnson, 1980). Societal attachment to certain terms and phrases also showcases how society shapes and is shaped by the power of language. For instance, Nunberg (2006) charts the shifting meaning of the word ‘liberal’ over time in American culture, noting how it has come to rest, most recently, as a word with not only negative, but primarily elitist connotations. His recent work (Nunberg, 2013) has also focused on words like ‘asshole’, which, when understood relative to their frequency, popularity, and emergent patterns, reveal correlate insights about their cultural context. As he shows, placing the common usage of a word within a larger socio-historical frame can mirror something about what binds a culture or subculture together as a people or animates shared beliefs or understandings. Language can showcase a momentary semantic resonance within a population, enlivening certain terms as cultural artifacts ripe for deeper social analysis.

This is all the more true when the terms are intentionally employed as hashtags. Social media, broadly writ, have been recognized by scholars and pundits alike as a reflection of culture (Abidin, 2014; Burgess and Green, 2009; Marwick, 2013; Murthy, 2018; Turner, 2013). Presidential agendas are leaked on Twitter, fashion trends are discovered on Instagram — yet one element of social media in particular stands out for its ability to crystallize cultural sentiment more than any other: the hashtag. On social media platforms, hashtags provide users with the ability to ascribe a post with meta-level information. Using a hashtag, one can highlight an important contextual element or signal an affiliation with a larger social conversation (Bruns and Burgess, 2011; Small, 2011). Set apart by the use of the # symbol, hashtags not only let a post’s author add this voluntary layer of meaning, but most importantly for scholars of culture, they allow researchers to categorize commonly hashtagged posts together (Darling, et al., 2013) and mine them for key contextual information or other relevant interpretive cues (Golder and Huberman, 2006; Marwick and boyd, 2011).

When collectively analyzed, hashtags can signal evidence of discussion around a specific topic (Bruns and Burgess, 2015; Zappavigna, 2012), event (i.e., emergency relief coordination) (Hughes and Palen, 2009), the development of protest networks (Hemsley and Eckert, 2014), signal the presence of a real or imagined community (Lorenzo-Dus and Di Cristofaro, 2016; Mendoza, et al., 2010), and even signal one’s identity (Tanupabrungsun and Hemsley, 2018). Moreover, since posts can contain multiple hashtags, both users and researchers can link these microsemantic cues together to surface the more complex meaning embedded in the complementary (or juxtaposed) connotations of a post’s co-occuring hashtags (Rieder, 2012). In short, analyzing hashtags and how they are used individually and in groups reveals an important dynamic regarding the way that cultural meaning is shaped (and reshaped) within social media (Blaszka, et al., 2012; Cunha, et al., 2011; Ma, et al., 2012) as well as by whom.

Our goal in this paper is to discover if there are clear groupings of terms that indicate micro-semantic spaces within the online discussion of mobile work. We believe that doing so will help us define different concepts within or around mobile work. As such, we propose the following research questions:

  1. Are there clear groupings of Twitter hashtags within and around mobile work?
  2. How do those groupings help us better understand the different dimensions of mobile work empirically?

Our selection of Twitter versus another social media platform for this study is based on two key criteria. First, Twitter acts like a type of news platform (Kwak, et al., 2010), making it a salient discourse space for tracking the breakout of new ideas and language. Second, Twitter is regarded as an influential cultural space. Twitter users skew towards elites like journalists, politicians, activists, and celebrities, thereby rendering the site as an important amplification medium; it can be said that Twitter users play a disproportionately high role in cultural agenda setting (Chadwick, 2013; Neuman, et al., 2014). Thus, terms that are present on Twitter can be considered not only potentially new, but also culturally significant.



3. Methods

3.1. Data

Data collection proceeded in two phases: a tag discovery phase, which lasted three months, and a 45-day formal data collection period. All data were collected from Twitter’s streaming API using an open source tool kit [1] that was chosen for its ease of use and its ability to collect Twitter rate limit notifications along with tweets that match a terms list, namely our list of hashtags.

In the first phase, the team started with an initial list of terms [2] (see the Appendix) based on existing research on mobile knowledge workers (Erickson and Jarrahi, 2016; Erickson, et al., 2014; Jarrahi and Thomson, 2017; Nelson, et al., 2017). This seed set of tags was expanded through tag chaining — that is, after the first month of tweet collection, the research team examined a random set of 3,000 of the total 141,421 tweets collected. This examination surfaced new, related hashtags that were previously unconsidered; it also identified a few terms that we deemed irrelevant or too general for our purposes. For example, we excluded the word “mobile” because it captured a vast array of unrelated tweets such as those related to mobile phones, the company Mobile Gas, Mobile Alabama, and so on. In general, however, we sought to be as inclusive as possible. After three iterations of this process, we settled on 51 hashtags.

Our formal collection period ran from 24 June 2016 to 8 August 2016, for a total of 45 days. We collected a total of 201,263 tweets, or an average of about 4,473 tweets per day. We received no rate limit notifications, which suggests we have an inclusive dataset. After removing retweets, we were left with 82,508, which we used for our analysis. These tweets contained a total of 18,399 different hashtags. As expected, many of our 51 search terms were among the most frequently occurring hashtags. For example, ‘coworking’ was one of our terms and is the most frequent tag in our set, showing up in 19,715 tweets. Table 1 shows the most frequently occurring hashtags in our set, whether they were part of our original list (bolded) or discovered due to co-occurrence (italicized), as well as the percentage of overall tweets they were found in.


Table 1: Top 20 hashtags. Search terms in bold, Co-occurring terms italicized.


A majority of the hashtags (10,725) show up in exactly one tweet in our dataset. The mean number of tweets a tag is present in is 12.56. Only 215 tags show up in more than 100 tweets, and only 30 tags occur in more than 1,000 tweets. There are 52,379 tweets (or 63 percent) with between two and four tags, and 11,732 (14 percent) have more than four tags, with the max being two tweets with 14 tags each.

It should be noted that the selection of relevant terms for the first round of tweet analysis was dependent on the interpretation of the authors, and partly shaped the outcomes. Thus, bias could be inherent in the data that shaped the outcome. However, we believe that the carefully considered selection of terms by experienced researchers helps to mitigate any severe bias. We are also confident that the iterative process we used surfaced most of the important terms related to mobile work in Twitter, especially given the scale of the data that we collected and the fact that many tweets contain multiple hashtags. In other words, if there were other terms in use, they ought to have surfaced as co-occurring tags with our list.

3.2. Analysis

To analyze this dataset, we used a complementary, mixed methodological approach that included computational work and some interpretive analysis. Computational techniques were used to identify patterns of hashtag use and surface interesting groupings in tweets. These groupings were then investigated qualitatively to help explain the patterns. We started by identifying the most frequently occurring hashtag pairs and then used network mapping and topic analysis to help identify topical groupings. Once we identified highly co-occurring pairs and groupings of hashtags, we used the tweet text as source data for interpretation. That is, we read tweets in the different groupings to ascertain the nature of the topical discussion in each. The computational work we did included co-occurrence identification, network mapping, and topic analysis for confirmation. Note that for the co-occurrence analysis and network mapping, only tweets with two or more hashtags were included (64,111); for the topic analysis all tweets were used.

Previous research indicates that semantically and conceptually similar hashtags are more likely to be mentioned in the same tweet (Jussila, et al., 2013; Rieder, 2012; Weng and Menczer, 2015). That is, the co-occurring of hashtags reflects a similar or joint meaning present in the tweet. Table 2 shows tag pairs [3] that appeared at least 1,000 times in the overall dataset, ranked by frequency.


Table 2: Top co-occurring hashtag pairs.
Hashtag pairFrequency


While hashtag pairs are useful for seeing what kinds of topics tend to be discussed together, they are less useful for seeing larger thematic structures. For that, we used network mapping. In our network, each node is a hashtag and each link represents a case where those tags were present in the same tweet. The resulting network contains 18,398 nodes and 92,335 links. Links are weighted by the number of times two hashtags co-occur in the same tweet. About 70 percent of links (64,178) have a weight of one, meaning that a pair of hashtags co-occur in only one tweet. In cases where a tweet contains a single tag, no new link is counted. Because we wanted to use network mapping to support ancillary qualitative interpretations, we chose to map our 51 search tags and the top 100 co-occurring tags. After further simplifying the network by dropping nodes with no links, we were left with a network of 138 nodes and 3,266 links.

To make the network plot (see Figure 1), we used the package iGraph (Csardi and Nepusz, 2006) in R with the Davidson and Harel (1996) layout algorithm. This force-directed method of layout calculation attempts to situate nodes closer together when the weights of their links is higher. The visual result is a depiction of different groupings of tags. Groups tend to form around the most used hashtags in the group and less frequent co-occurrences tend to be farther apart. With an understanding of the underlying algorithm, we can interpret the map as a visual artifact in a similar way one would interpret a photo.

In order to provide additional evidence to our visual interpretation, we performed topical modeling, which is a class of natural language processing that aims to discover sets of topics within a corpus of documents (Wang and McCallum, 2006; Blei and Lafferty, 2007; Ramage, et al., 2009; Rosen-Zvi, et al., 2004). Topics arise in the modeling process by uncovering clusters of related words. We used the MALLET [4] implementation of latent Dirichelt allocation (LDA) model on the text of our 82,508 tweets. As a generative statistical model, LDA assumes that each word in a corpus can be assigned to topics in a dataset with a different degree of membership (Karami, et al., 2019). Words with higher probability of membership in a given topic aggregate together, thereby uncovering otherwise hidden semantic meaning across a corpus of documents (Karami, et al., 2019). For this analysis, all stop words (e.g., “the”, “a”) were removed. LDA has been utilized to investigate different issues on Twitter, such as health (Sullivan, et al., 2016), politics (Karami and Elkouri, 2019), and disaster management (Caragea, et al., 2011).

Using MALLET allowed us to force the model iteratively into a limited number of topics (i.e., 25, 15, 5) to determine which topic number produced the best fit for the given corpus of text. Applying a trial and error approach (Chang, et al., 2009), we tried to identify the topic number with the smallest number of duplicates and the largest diversity. As we experimented, we found greater clarification and agreement with each successively smaller number of topics; utilizing more than three topics exhibited a high number of duplicates, whereas using three topics showcased an acceptable level of diversity. Ultimately, we found that limiting MALLET to three topics proved to be a highly parsimonious model.



4. Quantitative results


Hashtag co-occurrence network
Figure 1: Hashtag co-occurrence network: Nodes represent our collected hashtags and are colored depending on if they were in our search term list (blue) or were hashtags that happened to co-occur (brown). Nodes and labels are sized based on their frequency in the dataset. Links indicate that two terms co-occurred in the same tweet. The thickness of the link indicates how often those terms co-occurred. Blue links highlight two collected terms that co-occur in a tweet, while yellow links indicate that collected and incidental hashtags co-occur. Note also that since a tweet can have more than two hashtags, we can have a link between two tags that were incidentally collected.


Looking at the network map in Figure 1, we can see three distinct areas that form a triangle. The area at the top of the triangle is visually dominated by the hashtags digitalnomad, remotework, and travel. The second area (bottom right) is dominated by the hashtags coworking, workspace, nyc, and startup. The third area, (bottom left), is anchored with the hashtags hr, futureofwork, and flexibleworking. Notably one primary hashtag in our set — workanywhere — appears not to have a singular strong affiliation with any of these three areas, but instead forms a bridge among the other areas.

A comparison of these visual groupings with raw co-occurrence data provides additional nuance to build our understanding of the how the discourse space in Twitter is thematically partitioned. As indicated in Table 3, the first area, now called Area A, sees the tag digitalnomad overlapping most frequently with travel, followed closely behind by remotework and much less strongly with coworking. Likewise, remotework, the other central hashtag of this grouping, is most often linked with digitalnomad forming a tight symmetrical harmony; it overlaps less often, but nevertheless overlaps, with the tags futureofwork and digitalnomads [5] as well. Looking more closely at the numbers, we see that the tag workanywhere also links up with this semantic grouping, being most closely associated with the two core tags in this set.


Table 3: Hashtag frequencies and top co-occurrence counts.


Area B in the map, centered on coworking, is more complex than it may visually appear. Although there is an overlap between the hashtags coworking and workspace, it is not a forefronted co-occurrence. Two significant things are noteworthy, however, about this apparent, if loosely defined, area of discourse (which we expand on more fully below). First, coworking appears to be articulated as a platform-like space, something that is overtly linked to organizational types (i.e., startups) or worker types (i.e., digitalnomad). By contrast, workspace connotes physicality — and by all accounts primarily the physical environment of New York City or at least American soil most commonly.

In the final area of the network map, Area C, we see another loose set of tags. The tag futureofwork does not co-occur with the tag flexibleworking within its top ten co-occurrence dyads, but flexibleworking, by contrast, finds frequent pairing with futureofwork. Notably futureofwork also links fairly strongly with remotework, conceptually orienting it — at least in in some part — with the discussion(s) around digital nomads. Suffice to say that while strong delineations don’t seem to exist among them, we are confident that three different linguistic clusters appear to be present in the Twitter dataset.

The results of our topic analysis are provided in Table 4. Topic 1 approximates Area A with its association of digitalnomad, remotework, and futureofwork, while Topic 2 again resonates with Area B, particularly as it is visualized by the network algorithm. Finally, Topic 3 suggests a loose coupling of terms that correspond to the themes suggested by Area C in the network map and the co-occurrence statistics. As a confirmatory method, topic modeling helps to verify the three linguistic groupings we see in the map.


Table 4: Twitter dataset forced into three topics.
Topic 1digitalnomad, remotework, work, travel, remote, futureofwork, digitalnomads, future
Topic 2coworking, workspace, space, office, gt, startups, startup, spaces
Topic 3enterprisemobility, amp, working, remoteworking, business, mobile, flexibleworking, mobilefirst


Thus, to answer to our first research question, we can confirm that our multi-methodological approach reveals three distinct linguistic patterns related to mobile work in the Twitter posts we gathered. We might summarize alternately by saying that there appear to be three separate conversations, discursive spaces, trends, or themes present in Twitter related to mobile work. To minimize confusion moving forward, we elect to use the word ‘theme’ to refer to these linguistic patterns. The merits and meaning of these different themes, however, require further explanation.



5. Qualitative results and discussion

The first theme found in our quantitative analysis is represented in the data by network Area A and Topic 1. It centers on worker identity and is currently dominated by the experiences of digital nomads. The second theme is represented by Area B and Topic 2. This is a conversation about the practicalities of working somewhere physically and is currently dominated by issues related to co-working spaces. This theme appears to interweave the notion of space into workers’ sense of dislocation and mobility. Finally, the third theme correlates to Area C and Topic 3. It is a loose and speculative set of ideas around predicting how work will evolve in the future, with a particular emphasis on the role of the enterprise. We expand on each of these three themes below.

5.1. The evolution of the nomad

In organizational studies and related fields, early research that referenced the term ‘nomad’ elided any reference to the qualifier ‘digital’ (Ciolfi, et al., 2005; de Carvalho, et al., 2011; Kleinrock, 1995; Su and Mark, 2008). This is not so in today’s parlance. The tightest discursive cluster in our findings centers on the hashtag digitalnomad, which corresponds to a new, extreme form of location-independent work. Self-described digital nomads emphasize the balance between their work, personal life, and leisure time, often finding ways to be gainfully employed while traveling the world and hanging out on the beach (Dal Fiore, et al., 2014). Here, mobility is not merely part of the practice of a digital nomad, but part of the constitution of what it means to be a digital nomad. This thematic set of posts in Twitter principally reflects an act of sensemaking by workers and others around the newly emerging worker identity of digital nomad. The conversation reinforces the identification of this type of worker as one who is more than merely mobile, but rather is advantageously empowered by their mobility. The media these days (Chayka, 2018; Hart, 2015; Johanson, 2014) is redolent with stories of workers who engage in work that is either benefitted by their lack of fixedness or is not limited by it. This is not to say that digital nomads necessarily extol a nomadic way of life as superior, but more that they find ways to leverage their situation professionally.

Looking more closely at the tweets in this area (see Figure 2 for example), it is evident that two key sub-themes with regard to the experience of digital nomadism/nomadicity emerge from the data. The first theme centers on an exuberant pragmatics regarding digital nomads’ high level of mobility. Hashtags such as travel and adventure appear very frequently in these posts, but nearly as common is a direction for readers to attend to the tag ttot (Travel Talk on Twitter). The discussions that ensue there are not the standard fare of commuter rails and traffic jams, but rather involve details about new experiences and adventures surrounding world travel and leisure. Several popular locations for where digital nomads commune are mentioned via such hashtags as Thailand and London. In some particularly detailed tweets, Twitter posters will contribute links about alternative housing arrangements, such as local co-housing options.


A digital nomad tweets about work experiences
Figure 2: A digital nomad tweets about work experiences.


The second key sub-theme in these tweets corresponds to the type of workstyle arrangements commonly associated with digital nomadism — something quite distinct from the strategic discussions found in the first cluster of tweets. The co-occurring hashtags in these posts reveal a lot about how digital nomads describe and characterize their work/life arrangements: workhardanywhere, freedom, coliving, locationindependent. These coded phrases tend to have a positive spin, likely purposively underscoring both the agility that digital nomads think of themselves as having (i.e., the ability to be productive in any location) as well as the intentionality of this lifestyle (i.e., the free choice of living in such a manner). Embedded within these comments, however, we do not necessarily see a concomitant critique of the traditional 9a–5p, office-based lifestyle, from which we infer that these tweets are serving more to help digital nomads legitimate their own professional choices and resulting identities rather than disregard others.

This second sub-theme also contains hashtags like coliving, which are a reminder that work and life are tightly intertwined for digital nomads. Despite the fact that workers can conduct their work on the move, at some point they need to account for the basics of living. Thus for all the lack of fixedness in this population, we see a corresponding rise of fixed environments that cater to the merged professional and personal needs of digital nomads. A more nuanced consideration of the term ‘coliving’ might also reveal that digital nomads need new types of organizational bases with equally curious identities to account for and support their own emergent identity work. Because digital nomads are defined more by their practices than their professions in many cases, they often need to clarify that their work is still an elemental, if untraditionally defined, aspect of their identity. Much of this identity work appears to rest on the articulation of how their work and lifestyle is distinct and what advantages are embedded in this distinction. A common element here revolves around the benefits of work flexibility: flexibility in how, where and when they can engage with work and set the unique boundary between work and personal life (often defined by world travel adventure).

Moreover, the ethos of many of these tweets with their stress on making one’s life on one’s own terms might be described as entrepreneurial, so it is not surprising that hashtags related to entrepreneurship and startup are frequently used in conjunction with jobs and activities that lend themselves to location-independent nomadism. In this same vein, tweets in this thematic area frequently use a combination of remote plus another term to highlight digital nomads’ extreme dislocation. We feel that the use of these hashtag combinations are attempts to make clear that digital nomads’ activities diverge drastically from traditional conceptions of working remotely from home (Hinds and Bailey, 2003).

In sum, there is a recognizable theme in our tweets about mobile work that discuss how digital nomads conduct their work, find relevant communities and resources, leverage technological tools, maintain productivity, and divide their work and living spaces [6]. Most of these posts are directed at building identity among those who consider themselves digital nomads and persuading others about the virtues of digital nomadism.

5.2. Mobile in place

The second theme evident in our Twitter dataset is loosely organized by the hashtags coworking and workspace. This cluster of posts underscores the fact that work occurs in some particular location, however untethered and autonomous it may be appear to be. In other words, mobile workers — and even digital nomads — are not merely digital creatures zipping through fiber optic cables (though we assume many would warm to this idea); rather, these workers and their collaborators, as well as their devices and related infrastructures, are material and located in space. At some point a person needs to find an outlet to recharge her phone, for example. So, by contrast to the previous thematic cluster extolling the virtues of digital nomadicity, these posts are anchored by references to physical fixity — in specific, the places where mobile workers get their work done.

The most central hashtag in this cluster, coworking, highlights several important things about today’s population of mobile workers. First, many mobile workers appear to be finding what they need with regard to physical office space at coworking spaces, a physical office wherein independent workers may rent a desk, conference room, Wi-Fi and other facilities supporting work. Many posts in our dataset celebrate the connection between the worker and their coworking space in a way that is rarely seen by traditional office workers. We take this to mean, first, that these places adequately meet workers’ needs, but, secondly and potentially more importantly, mobile workers are relating to the places in which they choose to work with more than a passing regard. These relationships build on the physical characteristics of a space, as can be seen in the co-occurring hashtags workspace, officedesign, and officespace, as well as mention of their physical locations: nyc, usa, and manhattan. Workers are also emotional with regard to their work locations, which can be seen in the hashtag, lovewhereyouwork. These relations between worker and workspace are also increasingly dynamic, as can been seen by advertisement posts from coworking operations, like WeWork in particular, that see the rising population of digital nomads and other mobile workers as their main constituency. The large number of promotional tweets by coworking workspaces within our sample bears separate attention in future work to explore the terms of this dynamic relationship more deeply.

Analyzing the link between mobile work and coworking within this theme also helped to surface the entrepreneurial nature of much of today’s work life, according to those active on Twitter. The data suggests a weak link between coworking spaces and entrepreneurial, independent work, which can be seen in the co-occurring hashtags startup, startups, smallbusinessness, freelancers, and investing. These tweets also emphasize the vitality of co-working spaces as innovative communities.

Least evident in the data for this grouping, but worth mentioning even so, are the set of tweets that people post that are apparently recording their fixed experiences of being a mobile worker. Erickson (2008, 2007) refers to these kinds of posts as ‘self in place’ documentations, which serve to link an experience to a physical location. Secondarily this practice also allows a person to historically document their own mobility over time. The hashtags associated with this documentary practice, like the digital nomad tweets described earlier, emphasize how professional mobility is a type of lifestyle, complete with its own significant symbols (e.g., coffee). These posts are tagged with mobileofficelife, coffee, and appleandcoffee hashtags and are frequently accompanied with a photograph [7]. Beyond the utility of self-archiving, documentary posts like these might also be interpreted as reminders to distant collaborators of one’s existence, if otherwise unseen — a practice that has been well documented in earlier eras of distributed and remote work (Cramton, 2001; Erickson and Kellogg, 2000; Hinds and Bailey, 2003).

In sum, the second theme evident in our analysis of Twitter hashtags reminds us that mobile work has material dimensions and occurs in actual physical locations. These locations (i.e., coworking spaces), in turn, have cultural qualities that not only support particular kinds of work or professional attitudes well, such as entrepreneurial initiatives, but seem to engender a positive relationship between workers and their workspaces that is distinctive.

5.3. The future is flexible

The third theme we identified among the tweets we analyzed is loosely defined — or potentially nascent. It concerns the future of work (i.e., futureofwork hashtag). Given the topic, it is understandable that the boundaries around this theme are vague. That said, there are several dimensions that bear comment. First, most of the posts appear to be from an organizational point of view. This fact is evidenced by the first co-occurring tag, hr. Individuals do not have their own human resource departments (at least not yet), so already this link cues us into the fact that organizations may be trying to address issues surrounding mobile work to accord with their own policies and resources. Another correlated link, remotework, further connotes that knowledge workers will likely not be coming back to the traditional office workplace anytime soon; rather, working away from an office will be a trend into the future. However, the denotation of ‘remote’ assumes a centralized or core location, from which others are at a remove. Once again, one could read an organizational orientation into this reference. The co-occurring hashtag, ai (artificial intelligence), could again reveal an organizational emphasis, perhaps related to the eventual replacement of knowledge workers by machine learning or the management of mobile workers algorithmically. Alternately, it may be a remark on the fact that any conversation about the future of work — related to mobility or otherwise — is likely to address the issue of AI one way or the other. Additional analysis is needed to explicate the semantic relationships among these co-occurring terms in the data further.

Another tag present in this thematic cluster, enterprisemobility, yet again reinforces an organizational emphasis with a focus on the development and management of enterprise IT infrastructure in relation to changes in the mobility of the workforce and the dislocation of the workplace. Present in close proximity within the cluster is the related hashtag, mobilesecurity, which suggests that certain posters on Twitter are imagining the future security of mobile enterprise infrastructure and the consequential technological policies (e.g., byod[Bring Your Own Device]) that must accompany these decisions. While the details of these discussions are not yet fleshed out well in social media (nor perhaps would we expect them to be in this particular venue), the focus of the discourse is, again, evident as organizational.

Within the posts centered on the future of work in our dataset, we also see a dialogue centered on flexibleworking. When the co-occurring tags hiring and leading appear in tweets, we believe this refers to organizations’ concerns about the motivation and management of a mobile workforce. A large portion of the discourse is also constituted by the tags flexibleworking, smartworking, agileworking, and gigeconomy, and these read more like meta-level terms that HR departments (or media pundits) use as buzzwords to broadly paint new or imagined types of work practices without much regard to specifics. These references, as we noted at the beginning, likely showcase the population of journalists and the media actors who are using Twitter to discuss how mobility will comprise future forms of work. We might consider these posts as commentary by outsiders looking in, rather than mobile workers tweeting out. Contrast these tags to one from the digital nomad group, workhardanywhere, for example.

The veracity of this contention is supported by noting something about the co-occurrence of the tag futureofwork with the tags digitalnomad and digitalnomads. The singular tag (digitalnomad) is at the very far end of the cluster within our network map, while the plural tag is more central. We assert that this implies that there is a population in Twitter talking about digital nomads as a group of workers (digitalnomads), whereas the self-referential tweets (digitalnomad) are rarer within this theme — and, notably, more linked to the other clustered topics involving workspaces and the future of work. That said, Figure 2 uses the plural version of this tag, which may also indicate that there is a nascent community of practice growing among these workers. More data and analysis is required to claim anything definitive here.

In sum, this third theme in our Twitter data stresses an emerging conversation about mobile work within one of its largest discursive contexts: the discussion about how work may change in the future. Unlike the two other clusters that we detailed, this cluster is largely dominated by organizational and media voices opining on issues that are arising or will arise as work continues to evolve in the future. Like other predictions, this conversation is at best speculative, but nevertheless revealing in the manner in which terms collide in a freestyle form of semantic association and imagination.

5.4. Different dimensions of mobile work

Our second research question asks about different dimensions of mobile work. The three themes we identify suggest that historical terms such as distributed, remote and telework place a strong emphasis on workplace, but not on worker. This is in line with only one of our thematic discoveries — a theme that now appears to center on the concept of workplace more as an affordance of work than as a particularized context. Nomadic work of the past is now much more realized as a digitized phenomenon — movement is something that must be managed, as is one’s identity as a digital nomadic worker. Nomadicism is also seen in recent tweets much more as a choice than a problematic reality for a few types of workers.

Finally, both revealing but of no surprise is the fact that Twitter reflects the uncertainty of our times about where work is going and what it will look like when it gets there. This discussion, however, does not appear to come from the ranks of workers as much as it derives from organizations and commentators looking at the future of work as an object upon which they can or must speculate. For this reason, it appears to be organizationally oriented, focusing primarily on the general implications of changing norms of work and best practices for larger enterprises, tech firms, and decision-makers.



6. Conclusion

The three thematic clusters regarding the topic of mobile work we discovered in our Twitter dataset begin to extend our understanding of the experience of working on the move, identifying as a mobile worker, and preparing for a future where mobility is the common standard. Our study does present some limitations, however, which we would like to acknowledge. Using a singular type of data (i.e., tweets) should make it clear that we are not trying to claim a definitive understanding of mobile work, digital nomadism, enterprise security, or any of the other related topics that rose to our attention during our analysis. Rather, we think that using Twitter data is a means to conduct valuable exploratory research in fast-developing cultural areas. It allowed us to see that there are distinct thematic clusters in the data that reveal nascent trends ripe for further analysis. Such an approach is particularly valuable for directing future scholarly studies or refocusing the attention of organizations, the media, or the design world away from contemporary tropes or limited framings onto potentially more generative areas of consideration.

The growth in identification of one kind of worker in particular — the digital nomad — is not only redefining what it means to ‘work in place’, but also appears to be challenging the boundary between work and life itself by pursuing a professional life that is simultaneously a travel adventure. The distinctive role that coworking and coliving spaces are beginning to play in mobile workers’ lives reveals not only workers’ mobility ecosystems, but also the important role they play in providing support across a vast number of locations (Lee, et al., 2019; Gandini, 2016). These spaces have emerged as a type of the third place (Di Marino and Lapintie, 2017; Oldenburg and Brissett, 1982), facilitating nomads’ needs for knowledge sharing and community building. Yet they also serve as a locale through which they can hone their identities as work/life entrepreneurs. For these and likely other reasons, we can infer that digital nomads hold a dynamic and even paradoxical relationship with the notion of workplace — at once a place of support and identification while also being merely stop along an otherwise mobile path (Jarrahi, et al., 2019).

Finally, the beginnings of a speculative conversation about future forms of work in Twitter shows some early consideration of the potential effect of technologies on the nature of organizational roles, security, and human resource management. While the future of work may appear bright to digital nomads with their interests in independence and entrepreneurship, larger organizations and their constituents may be portending a more mixed future. If and when workers evolve into free agents, they will be not only more challenging to manage, but their relationships to their employers (i.e., transactional, distributed) may push against traditional forms of enterprise security organized around a fixed, internalized population.

To close, the three thematic clusters we identified in Twitter collectively represent an evolving balance between work practices, workers, and the broad array of sociotechnical infrastructures that support both. Social media references to digital nomads as an emerging and growing population of knowledge workers, co-working spaces as the new locales for knowledge work, and the speculative frontier regarding the future of work alert us to the fact that advancements and innovations are complex and are occurring on multiple fronts in numerous cultural locations. By analyzing everyday parlance on a popular social media channel like Twitter, we access a bit of this complexity. This forum of cultural tastemakers, sense-makers, and trendsetters offers us a glimpse of linguistic invention, robust identity making, and cautious reflection that we would do well to pay attention to if we want to embrace the potent intricacies of this topic with due rigor. End of article


About the authors

Jeff Hemsley is Assistant Professor at the School of Information Studies at Syracuse University.
E-mail: jjhemsle [at] syr [dot] edu

Ingrid Erickson is Assistant Professor at the School of Information Studies at Syracuse University.
E-mail: imericks [at] syr [dot] edu

Mohammad Hossein Jarrahi is Associate Professor in the School of Information and Library Science at the University of North Carolina at Chapel Hill.
E-mail: jarrahi [at] unc [dot] edu

Amir Karami is Assistant Professor in the School of Library and Information Science at the University of South Carolina.
E-mail: karami [at] mailbox [dot] sc [dot] edu




2. See the Appendix for the terms list as well as more detail on the process.

3. The hashtags sks8 and sneek refer to commercial organizations so these pairs are primarily self-promotional. The former is a marketing agency and the latter is a startup that makes a platform that supports distributed teams.


5. Note that in this work we do not stem terms, so digitalnomad and digitalnomads are separate hashtag search terms. For example, a person may self-identify as a digitalnomad in Twitter, while someone else market to a perceived audience of digitialnomads. We believe that including both versions of this tag provides nuance for our analysis.

6. More on this topic can be found here:

7. Figure 2 would also fit this type of post.



Crystal Abidin, 2014. “#In$tagLam: Instagram as a repository of taste, a burgeoning marketplace, a war of eyeballs,” In: Marsha Berry and Max Schleser (editors). Mobile media making in an age of smartphones. New York: Palgrave Pivot, pp. 119–128.
doi:, accessed 14 February 2020.

Diane E. Bailey, Stephanie Layne Dailey, Paul Leonardi, Bonnie Nardi, and Eduardo Henrique Diniz, 2015. “Socializing remote workers: Identification and role innovation at a distance,” Academy of Management Proceedings, volume 15, number 1.
doi:, accessed 14 February 2020.

Matthew Blaszka, Lauren M. Burch, Evan L. Frederick, Galen Clavio, and Patrick Walsh, 2012. “#WorldSeries: An empirical examination of a Twitter hashtag during a major sporting event,” International Journal of Sport Communication, volume 5, number 4, pp. 435–453.
doi:, accessed 14 February 2020.

David M. Blei and John D. Lafferty, 2007. “A correlated topic model of Science,” Annals of Applied Statistics, volume 1, number 1, pp. 17–35.
doi:, accessed 14 February 2020.

Nicola Bozzi, in press. “#digitalnomads, #solotravellers, #remoteworkers: A cultural critique of the travelling entrepreneur on Instagram,” Social Media + Society; version available at, accessed 14 February 2020.

Axel Bruns and Jean Burgess, 2015. “Twitter hashtags from ad hoc to calculated publics,” In: Nathan Rambukkana (editor). Hashtag publics: The power and politics of discursive networks. New York: Peter Lang, pp. 13–28.
doi:, accessed 14 February 2020.

Axel Bruns and Jean Burgess, 2011. “The use of Twitter hashtags in the formation of ad hoc publics,” Proceedings of the Sixth European Consortium for Political Research (ECPR) General Conference 2011, at, accessed 9 September 2016.

Jean Burgess and Joshua Green, 2009. YouTube: Online video and participatory culture. Cambridge: Polity.

Cornelia Caragea, Nathan McNeese, Anuj Jaiswal, Greg Traylor, Hyun-Woo Kim, Prasenjit Mitra, Dinghao Wu, Andrea H. Tapia, Lee Giles, Bernard J. Jansen, and John Yen, 2011. “Classifying text messages for the Haiti earthquake,” Proceedings of the Eighth International ISCRAM Conference, at, accessed 14 February 2020.

James W. Carey, 2008. Communication as culture: Essays on media and society. Revised edition. New York: Routledge.

Anthony Chadwick, 2013. The hybrid media system: Politics and power. New York: Oxford University Press.
doi:, accessed 14 February 2020.

Jonathan Chang, Sean Gerrish, Chong Wang, Jordan L. Boyd-Graber, and David M. Blei, 2009. “Reading tea leaves: How humans interpret topic models,” Advances in Neural Information Processing Systems, pp. 288–296, and at, accessed 14 February 2020.

Maria Charalampous, Christine A. Grant, Carlo Tramontano, and Evie Michailidis, 2019. “Systematically reviewing remote e-workers’ well-being at work: A multidimensional approach,” European Journal of Work and Organizational Psychology, volume 28, number 1, pp. 51–73.
doi:, accessed 14 February 2020.

Kyle Chayka, 2018. “When you’re a ‘digital nomad,’ the world is your office,” New York Times (8 February), at, accessed 16 November 2019.

Luigina Ciolfi and Aparecido Fabiano Pinatti de Carvalho, 2014. “Work practices, nomadicity and the mediational role of technology,” Computer Supported Cooperative Work, volume 23, number 2, pp. 119–136.
doi:, accessed 14 February 2020.

Luigina Ciolfi, Iride Bartolucci, and Darragh Murphy, 2005. “Meaningful interactions for meaningful places: Investigating the relationships between nomadic work, tangible artefacts and the physical environment,” EACE ’05: Proceedings of the 2005 Annual Conference on European Association of Cognitive Ergonomics, pp. 115–121.

Jana Costas, 2013. “Problematizing mobility: A metaphor of stickiness, non-places and the kinetic elite,” Organization Studies, volume 34, number 10, pp. 1,467–1,485.
doi:, accessed 14 February 2020.

Karlene Cousins and Daniel Robey, 2015. “Managing work-life boundaries with mobile technologies: An interpretive study of mobile work practices,” Information Technology & People, volume 28, number 1, pp. 34–71.
doi:, accessed 14 February 2020.

Catherine Durnell Cramton, 2001. “The mutual knowledge problem and its consequences for dispersed collaboration,” Organization Science, volume 12, number 3, pp. 346–371.
doi:, accessed 14 February 2020.

Gabor Csardi and Tamas Nepusz, 2006. “The igraph software package for complex network research,” InterJournal, at, accessed 14 February 2020.

Evandro Cunha, Gabriel Magno, Giovanni Comarela, Virgilio Almeida, Marcos André Gonçalves, and Fabrício Benevenuto, 2011. “Analyzing the dynamic evolution of hashtags on Twitter: A language-based approach,” LSM ’11: Proceedings of the Workshop on Languages in Social Media, pp. 58–65.

Filippo Dal Fiore, Patricia L. Mokhtarian, Ilan Salomon, and Matan E. Singer, 2014. “‘Nomads at last’? A set of perspectives on how mobile technology may affect travel,” Journal of Transport Geography, volume 41, pp. 97–106.
doi:, accessed 14 February 2020.

Emily S. Darling, David Shiffman, Isabelle M. Côté, and Joshua A. Drew, 2013. “The role of Twitter in the life cycle of a scientific publication,” Ideas in Ecology and Evolution, volume 6, pp. 32–43.
doi:, accessed 14 February 2020.

Ron Davidson and David Harel, 1996. “Drawing graphs nicely using simulated annealing,” ACM Transactions on Graphics, volume 15, number 4, pp. 301–331.
doi:, accessed 14 February 2020.

Aparecido Fabiano Pinatti de Carvalho, Luigina Ciolfi, and Breda Gray, 2017. “Detailing a spectrum of motivational forces shaping nomadic practices,” CSCW’17: Proceedings of the 2017 ACM Conference on Computer Supported Cooperative Work and Social Computing, pp. 962a–977.
doi:, accessed 14 February 2020.

Aparecido Fabiano Pinatti de Carvalho, Luigina Ciolfi, and Breda Gray, 2011. “The making of nomadic work: Understanding the mediational role of ICTs,” In: M.M. Cruz-Cunha and F. Moreira (editors). Handbook of research on mobility and computing: Evolving technologies and ubiquitous impacts. Hershey, Pa.: IGI Global, volume 1, pp. 381–396.
doi:, accessed 14 February 2020.

Mina Di Marino and Kimmo Lapintie, 2017. “Emerging workplaces in post-functionalist cities,” Journal of Urban Technology, volume 24, number 3, pp. 5–25.
doi:, accessed 14 February 2020.

Alan J. Dix and Russell Beale (editors), 1996. Remote cooperation: CSCW issues for mobile and teleworkers. Berlin: Springer.
doi:, accessed 14 February 2020.

Nicola Ens, Mari-Klara Stein, and Tina Blegind Jensen, 2018. “Decent digital work: Technology affordances and constraints,” ICIS 2018 Proceedings, at, accessed 14 February 2020.

Michael Erard, 2016. “Remote? That’s no way to describe this work,” New York Times (18 June), at, accessed 20 June 2016.

Ingrid Erickson, 2008. “The translucence of Twitter,” Ethnographic Praxis in Industry, volume 2008, number 1, pp. 64–78.
doi:, accessed 14 February 2020.

Ingrid Erickson, 2007. “Understanding socio-locative practices,” GROUP’07 Doctoral Consortium papers, article number 1, pp. 1–2.
doi:, accessed 14 February 2020.

Ingrid Erickson and Mohammad Hossein Jarrahi, 2016. “Infrastructuring and the challenge of dynamic seams in mobile knowledge work,” CSCW ’16: Proceedings of the 19th ACM Conference on Computer-Supported Cooperative Work & Social Computing, pp. 1,323–1,336.
doi:, accessed 14 February 2020.

Ingrid Erickson, Mohammad H. Jarrahi, Leslie Thomson, and Steve Sawyer, 2014. “More than nomads: Mobility, knowledge work, and infrastructure,” EGOS 2014, at, accessed 14 February 2020.

Thomas Erickson and Wendy A. Kellogg, 2000. “Social translucence: An approach to designing systems that support social processes,” ACM Transactions on Computer-Human Interaction, volume 7, number 1, pp. 59–83.
doi:, accessed 14 February 2020.

Rosemary Francisco and Amarolinda Klein, 2018. “Knowledge workers on the move: Understanding how they use mobile ICT to create and share Knowledge,” Academy of Management Proceedings, volume 2018, number 1.
doi:, accessed 14 February 2020.

Alessandro Gandini, 2016. “Coworking: The freelance mode of organisation?” In: Alessandro Gandini. The reputation economy. London: Palgrave Macmillan, pp. 97–105.
doi:, accessed 14 February 2020.

James Gee, 2015. Social linguistics and literacies: Ideology in discourses. Fifth edition. London: Routledge.
doi:, accessed 14 February 2020.

Scott A. Golder and Bernardo A. Huberman, 2006. “Usage patterns of collaborative tagging systems,” Journal of Information Science, volume 32, number 2, pp. 198–208.
doi:, accessed 14 February 2020.

Ulrike Gretzel and Anne Hardy, 2019. “#VanLife: Materiality, makeovers and mobility amongst digital nomads,” e-Review of Tourism Research, volume 16, numbers 2–3, pp. 1–9, at, accessed 14 February 2020.

Brian M. Harmer and David J. Pauleen, 2012. “Attitude, aptitude, ability and autonomy: The emergence of ‘offroaders’, a special class of nomadic worker,” Behaviour & Information Technology, volume 31, number 5, pp. 439–451.
doi:, accessed 14 February 2020.

Anna Hart, 2015. “Living and working in paradise: The rise of the ‘digital nomad’,” Daily Telegraph (17 May), at, accessed 18 September 2016.

Jeff Hemsley and Josef Eckert, 2014. “Examining the role of ‘place’ in Twitter networks through the lens of contentious politics,” 2014 47th Hawaii International Conference on System Sciences, pp. 1,844–1,853.
doi:, accessed 14 February 2020.

Pamela J. Hinds and Diane E. Bailey, 2003. “Out of sight, out of sync: Understanding conflict in distributed teams,” Organization Science, volume 14, number 6, pp. 615–632.
doi:, accessed 14 February 2020.

Amanda Lee Hughes and Leysia Palen, 2009. “Twitter adoption and use in mass convergence and emergency events,” International Journal of Emergency Management, volume 6, numbers 3–4, pp. 248–260.
doi:, accessed 14 February 2020.

Mohammad Hossein Jarrahi and Leslie Thomson, 2017. “The interplay between information practices and information context: The case of mobile knowledge workers,” Journal of the Association for Information Science and Technology, volume 68, number 5, pp. 1,073–1,089.
doi:, accessed 14 February 2020.

Mohammad Hossein Jarrahi, Gabriela Philips, Will Sutherland, Steve Sawyer, and Ingrid Erickson, 2019. “Personalization of knowledge, personal knowledge ecology, and digital nomadism,” Journal of the Association for Information Science and Technology, volume 70, number 4, pp. 313–324.
doi:, accessed 14 February 2020.

Mark Johanson, 2014. “For digital nomads, work is no longer a place and life is one big adventure,” International Business Times (25 March), at, accessed 3 June 2017.

Jari Jussila, Jukka Huhtamäki, Hannu Kärkkäinen, and Kaisa Still, 2013. “Information visualization of Twitter data for co-organizing conferences,” AcademicMindTrek ’13: Proceedings of International Conference on Making Sense of Converging Mediapp, 139–145.
doi:, accessed 14 February 2020.

Amir Karami and Aida Elkouri, 2019. “Political popularity analysis in social media,” In: Natalie Greene Taylor, Caitlin Christian-Lamb, Michelle H. Martin, and Bonnie Nardi (editors). Information in contemporary society. Lecture Notes in Computer Science, volume 11420. Cham, Switzerland: Springer, pp. 456–465.
doi:, accessed 14 February 2020.

Amir Karami, Suzanne C. Swan, Cynthia Nicole White, and Kayla Ford, 2019. “Hidden in plain sight for too long: Using text mining techniques to shine a light on workplace sexism and sexual harassment,” arXiv (1 July), at, accessed 14 February 2020.

Leonard Kleinrock, 1995. “Nomadic computing — An opportunity,” ACM SIGCOMM Computer Communication Review, volume 25, number 1, pp. 36–40.
doi:, accessed 14 February 2020.

Benjamin Koehne, Patrick C. Shih, and Judith S. Olson, 2012. “Remote and alone: Coping with being the remote member on the team,” CSCW ’12: Proceedings of the ACM 2012 Conference on Computer Supported Cooperative Work, pp. 1,257–1,266.
doi:, accessed 14 February 2020.

Johanna Koroma, Ursula Hyrkkänen, and Matti Vartiainen, 2014. “Looking for people, places and connections: Hindrances when working in multiple locations: A review,” New Technology, Work and Employment, volume 29, number 2, pp. 139–159.
doi:, accessed 14 February 2020.

Haewoon Kwak, Changhyun Lee, Hosung Park, and Sue Moon, 2010. “What is Twitter, a social network or a news media?” WWW ’10: Proceedings of the 19th International Conference on World Wide Web, pp. 591–600.
doi:, accessed 14 February 2020.

George Lakoff and Mark Johnson, 1980. “Conceptual metaphor in everyday language,” Journal of Philosophy, volume 77, number 8, pp. 453–486.
doi:, accessed 14 February 2020.

Aurelie Leclercq-Vandelannoitte and Henri Isaac, 2016. “The new office: How coworking changes the work concept,” Journal of Business Strategy, volume 37, number 6, pp. 3–9.
doi:, accessed 14 February 2020.

Ahreum Lee, Austin L. Toombs, and Ingrid Erickson, 2019. “Infrastructure vs. community: Co-spaces confront digital nomads’ paradoxical needs,” PCHI EA ’19: Extended Abstracts of the 2019 CHI Conference on Human Factors in Computing Systems, paper number LBW2419, pp. 1–6.
doi:, accessed 14 February 2020.

Nuria Lorenzo-Dus and Matteo Di Cristofaro, 2016. “#Living/minimum wage: Influential citizen talk in Twitter,” Discourse, Context & Media, volume 13, part A, pp. 40–50.
doi:, accessed 14 February 2020.

Zongyang Ma, Aixin Sun, and Gao Cong, 2012. “Will this# hashtag be popular tomorrow?” SIGIR ’12: Proceedings of the 35th International ACM SIGIR Conference on Research and development in Information Retrieval, pp. 1,173–1,174.
doi:, accessed 14 February 2020.

Alice E. Marwick, 2013. Status update: Celebrity, publicity, and branding in the social media age. New Haven, Conn.: Yale University Press.

Alice E. Marwick and danah boyd, 2011. “I tweet honestly, I tweet passionately: Twitter users, context collapse, and the imagined audience,” New Media & Society, volume 13, number 1, pp. 114–133.
doi:, accessed 14 February 2020.

Marcelo Mendoza, Barbara Poblete, and Carlos Castillo, 2010. “Twitter under crisis: Can we trust what we RT?” SOMA ’10: Proceedings of the First Workshop on Social Media Analytics, pp. 71–79.
doi:, accessed 14 February 2020.

Dhiraj Murthy, 2018. Twitter. Second edition. Cambridge: Polity.

Caleece Nash, Mohammad Hossein Jarrahi, Will Sutherland, and Gabriela Phillips, 2018. “Digital nomads beyond the buzzword: Defining digital nomadic work and use of digital technologies,” In: Gobinda Chowdhury, Julie McLeod, Val Gillet, and Peter Willett (editors). Transforming digital worlds. Lecture Notes in Computer Science, volume 10766. Cham, Switzerland: Springer, pp. 207–217.
doi:, accessed 14 February 2020.

Sarah Beth Nelson, Mohammad Hossein Jarrahi, and Leslie Thomson, 2017. “Mobility of knowledge work and affordances of digital technologies,” International Journal of Information Management, volume 37, number 2, pp. 54–62.
doi:, accessed 14 February 2020.

W. Russell Neuman, Lauren Guggenheim, S. Mo Jang, and Soo Young Bae, 2014. “The dynamics of public attention: Agenda-setting theory meets big data,” Journal of Communication, volume 64, number 2, pp. 193–214.
doi:, accessed 14 February 2020.

Geoffrey Nunberg, 2013. Ascent of the a-word: Assholism, the first sixty years. New York: PublicAffairs.

Geoffrey Nunberg, 2006. Talking right: How conservatives turned liberalism into a tax-raising, latte-drinking, sushi-eating, Volvo-driving, New York times-reading, body-piercing, Hollywood-loving, left-wing freak show. New York: PublicAffairs.

Satu Ojala and Pasi Pyöriä, 2018. “Mobile knowledge workers and traditional mobile workers: Assessing the prevalence of multi-locational work in Europe,” Acta Sociologica, volume 61, number 4, pp. 402–418.
doi:, accessed 14 February 2020.

Ramon Oldenburg and Dennis Brissett, 1982. “The third place,” Qualitative Sociology, volume 5, number 4, pp. 265–284.
doi:, accessed 14 February 2020.

Marko Orel, 2019. “Coworking environments and digital nomadism: Balancing work and leisure whilst on the move,” World Leisure Journal, volume 61, number 3, pp. 215–227.
doi:, accessed 14 February 2020.

Wanda J. Orlikowski and Susan V. Scott, 2016. “Digital work: A research agenda,” In: Barbara Czarniawska (editor). A research agenda for management and organization studies. Northampton Mass.: Edward Elgar, pp. 88–95.
doi:, accessed 14 February 2020.

Daniel Ramage, David Hall, Ramesh Nallapati, and Christopher D. Manning, 2009. “Labeled LDA: A supervised topic model for credit attribution in multi-labeled corpora,” EMNLP ’09: Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing, volume 1, pp. 248–256.

Ina Reichenberger, 2018. “Digital nomads — A quest for holistic freedom in work and leisure,” Annals of Leisure Research, volume 21, number 3, pp. 364–380.
doi:, accessed 14 February 2020.

Bernhard Rieder, 2012. “The refraction chamber: Twitter as sphere and network,” First Monday, volume 17, number 11, at, accessed 2 September 2016.
doi:, accessed 14 February 2020.

Kevin W. Rockmann and Michael G. Pratt, 2015. “Contagious offsite work and the lonely office: The unintended consequences of distributed work,” Academy of Management Discoveries, volume 1, number 2, pp. 150–164.
doi:, accessed 14 February 2020.

Michal Rosen-Zvi, Thomas Griffiths, Mark Steyvers, and Padhraic Smyth, 2004. “The author-topic model for authors and documents,” UAI ’04: Proceedings of the 20th Conference on Uncertainty in Artificial Intelligence, pp. 487–494.

Julia Schlegelmilch and Evgenia Lysova, 2018. “Digital nomads and the future of work,” Academy of Management Proceedings, volume 2018, number 1.
doi:, accessed 14 February 2020.

Kjeld Schmidt and Liam Bannon, 1992. “Taking CSCW seriously,” Computer Supported Cooperative Work, volume 1, numbers 1–2, pp. 7–40.
doi:, accessed 14 February 2020.

Tamara A. Small, 2011. “What the hashtag? A content analysis of Canadian politics on Twitter,” Information, Communication & Society, volume 14, number 6, pp. 872–895.
doi:, accessed 14 February 2020.

Carsten Sørensen, 2011. Enterprise mobility: Tiny technology with global impact on work. London: Palgrave Macmillan.
doi:, accessed 14 February 2020.

Rosie Spinks, 2015. “Meet the ‘digital nomads’ who travel the world in search of fast wi-fi,” Guardian (16 June), at, accessed 18 September 2016.

Clay Spinuzzi, Zlatko Bodrožić, Giuseppe Scaratti, and Silvia Ivaldi, 2019. “‘Coworking is about community’: But what is ‘community’ in coworking?” Journal of Business and Technical Communication, volume 33, number 2, pp. 112–140.
doi:, accessed 14 February 2020.

Norman Makoto Su and Gloria Mark, 2008. “Designing for nomadic work,” >DIS ’08: Proceedings of the Seventh ACM Conference on Designing Interactive Systems, pp. 305–314.
doi:, accessed 14 February 2020.

Ryan Sullivan, Abeed Sarker, Karen O’Connor, Amanda Goodin, Mark Karlsrud, and Graciela Gonzalez, 2016. “Finding potentially unsafe nutritional supplements from user reviews with topic modeling,” Biocomputing 2016: Proceedings of the Pacific Symposium, pp. 528–539.
doi:, accessed 14 February 2020.

Will Sutherland and Mohammad Hossein Jarrahi, 2017. “The gig economy and information infrastructure: The case of the digital nomad community,” Proceedings of the ACM on Human-Computer Interaction, article number 97.
doi:, accessed 14 February 2020.

Sikana Tanupabrungsun and Jeff Hemsley, 2018. “Studying celebrity practices on Twitter using a framework for measuring media richness,” Social Media + Society (15 March).
doi:, accessed 14 February 2020.

Beverly Yuen Thompson, 2019. “The digital nomad lifestyle: (Remote) work/leisure balance, privilege, and constructed community,” International Journal of the Sociology of Leisure, volume 2, numbers 1–2, pp. 27–42.
doi:, accessed 14 February 2020.

Leslie Thomson and Mohammad Hossein Jarrahi, 2015. “Information practices in the broader ‘deportment’ of mobile knowledge work,” Proceedings of the Association for Information Science and Technology, volume 52, number 1, pp. 1–10.
doi:, accessed 14 February 2020.

Graeme Turner, 2013. Understanding celebrity. Second edition. Thousand Oaks, Calif.: Sage.

Xuerui Wang and Andrew McCallum, 2006. “Topics over time: A non-Markov continuous-time model of topical trends,” KDD ’06: Proceedings of the 12th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 424–433.
doi:, accessed 14 February 2020.

Lilian Weng and Filippo Menczer, 2015. “Topicality and impact in social media: Diverse messages, focused messengers,” PLoS ONE, volume 10, number 2, e0118410.
doi:, accessed 14 February 2020.

Alex J. Wood, Vili Lehdonvirta, and Mark Graham, 2018. “Workers of the Internet unite? Online freelancer organisation among remote gig economy workers in six Asian and African countries,” New Technology, Work and Employment, volume 33, number 2, pp. 95–112.
doi:, accessed 14 February 2020.

Michele Zappavigna, 2012. Discourse of Twitter and social media: How we use language to create affiliation on the Web. New York: Continuum.

Shoshana Zuboff, 1988. In the age of the smart machine: The future of work and power. New York: Basic Books.




We elected to select our initial tweet sample based on a set of terms that were semantically related to the broad topic of mobility and knowledge work. The second and third author generated these tags based on prior research with mobile knowledge workers (Erickson and Jarrahi, 2016; Erickson, et al., 2014) (see round 1 in the table below). The terms in this list were used as search terms for tweet collection from Twitter’s API. After one month, the authors met and examined example tweets to determine if the returned tweets were relevant to the overarching topic of forms of mobile work. When we agreed that a term was too noisy (the majority of tweets were off topic), or was clearly promotional (we removed the related hashtag WeWork, which is the name of a popular international coworking enterprise. The presence of promotional tweets for this organization was so predominant in the tweet results that we decided that it was best to bracket it for separate analysis at a later time), we removed the term from the next collection iteration. In general, however, we sought to expand the discussion space and so added 27 new hashtags that were determined to be consistently relevant. This gave us a total of 55 search terms for our second round of data collection, yielding a much larger incoming stream of tweets over the next month: 6,594,202.

The expanded set of search terms engendered a broad conversational space to examine, to be sure, but also amplified the noise. We engaged in one more round of culling prompted by a second round of close reading sample tweets. Intriguingly, our prior inclusion of general terms like mobile in the second data collection round produced a great deal of noise in the form of very loosely related tweets (e.g., a link to a news story about traffic accidents due to mobile phones). In the end, our overall, iterative data collection process resulted in a total corpus of 16,325,238 tweets, collected between 27 April 2016 and 8 August 2016.

To get our final data set for analysis, we selected a subset from the whole based on a set of filtering rules. First, since the final change to our collection terms set was 24 June 2016, we only included tweets after that date. Second, to reduce the problem of noise, we select only tweets with hashtag versions of our terms. We also opted to focus on hashtags made up of concatenated words since these seemed to be deliberate attempts to categorize content. Hashtags like #MobileOffice, #MobileWorkers, and #MobileWorking were kept, while #Mobile was let go.


round 1round 2round 3



Editorial history

Received 19 August 2019; revised 20 November 2019; revised 30 January 2020; accepted 1 February 2020.

Copyright © 2020, Jeff Hemsley, Ingrid Erickson, Mohammad Hossein Jarrahi, and Amir Karami. All Rights Reserved.

Digital nomads, coworking, and other expressions of mobile work on Twitter
by Jeff Hemsley, Ingrid Erickson, Mohammad Hossein Jarrahi, and Amir Karami.
First Monday, Volume 25, Number 3 - 2 March 2020