One of the central issues associated with altmetrics (short for alternative metrics) is the identification of communities engaging with scholarly content on social media (Haustein, Bowman, & Costas, 2015; Neylon, 2014; Tsou, Bowman, Ghazinejad, & Sugimoto, 2015) . It is thus of central importance to understand the uses and users of social media in the context of scholarly communication.

most identify the following major categori es: social networking, social bookmarking, blogging, microblogging, wikis , and media and data sharing (Gu & Widén -Wulff, 2011; Rowlands, Nicholas, Russell, Canty, & Watkinson, 2011; Tenopir et al., 2013) . Some also conside r conferencing, collaborative authoring, scheduling and meeting tools (Rowlands et al., 2011) or RSS and online documents (Gu & Widén -Wulff, 2011; Tenopir et al., 2013) as social media. The landscape of social media, as well as that of altmetrics, is constantly changing and boundaries with othe r online platforms and traditional metrics are fuzzy. Many online platforms cannot be easily classified and more traditional metrics , such as downloads and mentions in policy documents , have been referred to as altmetrics due to data pr ovider policies.

the Use of social media platforms for by researchers is high — ranging from 75 to 80% in large -scale surveys (Rowlands et al., 2011; Tenopir et al., 2013; Van Eperen & Marincola, 2011) .

but

less than 10% of scholars reported using Twitter (Rowlands et al., 2011) , while 46% used ResearchGate (Van Noorden, 2014) , and more than 55% use d YouTube (Tenopir et al., 2013) —it is necessary to discuss the use of various types of social media separately . Furthermore, there i s a distinction among types of us e, with studies showing higher uses of social media for dissemination, consumption, communication , and promotion (e.g., Arcila -Calderón, Piñuel -Raigada, & Calderín -Cruz, 2013; Van Noorden, 2014) , and fewer instances of use for creation (i.e., using social media to construct scholarship) (British Library et al., 2012; Carpenter, Wetheridge, Tanner, & Smith, 2012; Procter et al., 2010b; Tenopir et al., 2013) .

Frequently mentioned social platforms in scholarly communication research include research -specific tools such as Mendeley, Zotero, CiteULike, BibSonomy, and Connotea (now defunct) as well as general tools such as Delicious and Digg (Hammond, Hannay, Lund, & Scott, 2005; Hull, Pettifer, & Kell, 2008; Priem & Hemminger, 2010; Reher & Haustein, 2010) .

Social data sharing platforms provide an infrastructure to share various types of scholarly objects —including datasets, software code, figures, presentation slides and videos —and for users to interact with these objects (e.g., comment on, favorite, like , and reuse ). Platforms such as Figshare and SlideShare disseminate scholars’ various types of research outputs such as datasets, figures, infographics, documents, videos, posters , or presentation slides (Enis, 2013) and displays views, likes, and shares by other users (Mas -Bleda et al., 2014) . GitHub provides for uploading and stor ing of software code, which allows users to modify and expand existing code (Dabbish, Stuart, Tsay, & Herbsleb, 2012) , which has been shown to lead to enhanced collaboratio n among developers (Thung, Bissyande, Lo, & Jiang, 2013) . As w ith other social data sharing platforms, usage statistics on the number of view and contributions to a project are provided (Kubilius, 2014) . The registry of research data repositories, re3data.org, ha s indexed more than 1,200 as of May 2015 2 . However, only a few of these repositories (i.e. , Figshare, SlideShare and Github) include social functionalities and have reached a certain level of participation from scholars (e.g., Begel, Bosch, & Storey, 2013; Kubilius, 2014) .

Video provide s yet another genre for social interaction and scholarly communication (Kousha, Thelwall, & Abdoli, 2012; Sugimoto & Thelwall, 2013) . Of the various video sharing platforms, YouTube, launched in 2005, is by far the most popular

A study of UK scholars reports that the majority o f respondents engaged with video for scholarly communication purposes (Tenopir et al., 2013) , yet only 20% have ever created in that genre. Among British PhD students, 17% had used videos and podcasts passively for research, while 8% had actively contributed (British Library et al., 2012) .

Blogs began in the mid -1990s and were considered ubiquitous by the mid- 200 0s (Gillmor, 2006; Hank, 2011; Lenhart & Fox, 2006; Rainie, 2005) . Scholarly blogs emerged during this time with their own neologisms (e.g., blogademia , blawgosphere , bloggership) and body of research (Hank, 2011) and were considered to change the exclusive structure of scholarly communication

Technorati, considered t o be on e of the largest ind ex of blogs, deleted their entire blog directory in 2014 3 . Individual blogs are also subject to abrupt cancellations and deletions, making questionable the degree to which blogging meets the permanence criteria of scholarly commu nication (Hank, 2011) .

ResearchBlogging.org (RB) — “an aggregator of blog posts referencing peer -reviewed research in a structured manner” (Shema, Bar -Ilan, & Thelwall, 2015, p. 3) — was launched in 2007 and has been a fairly stable structure in the scholarly blogging environment. RB both aggregates and —through the use of the RB icon — credentials scholarly blogs (Shema et al., 2015) . The informality of the genre (Mewburn & Thomson, 2013) and the ability to circumve nt traditional publishing barr iers has led advocates to claim that blogging can invert traditional academic power hierarchies (Walker, 2006) , allow ing people to construct scholarly identities outside of formal institutionalization (Ewins, 2005; Luzón, 2011; Potter, 2012) and democratize the scientific system (Gijón, 2013) . Another positive characteristic of blogs is their “inherently social” nature (Walker, 2006, p. 132) (see also Kjellberg, 2010; Luzón, 2011 ). Scholars have noted the potential for “communal scholarship” (Hendrick, 2012) made by linking and commenting, calling the platform “a new ‘third place’ for academic discourse” (Halavais, 2006, p. 117) . Commenting functionalities were seen as making possible the “shift from public understanding to public engagement with science” (Kouper, 2010, p. 1) .

Studies have also provided evidence of high rate s of blogging among certain subpopulations: for example, approximately one -third of German university staff (Pscheida et al., 2013) and one fifth of UK doctoral students use blogs (Carpenter et al., 2012) .

while there has been anecdotal evidence of the use of blogs in promotion and tenure (e.g., (Podgor, 2006) the consensus seem s to suggest that most institutions do not value blogging as highly as publishing in traditional outlets, or consider blogging as a measure of service rather than research activity (Hendricks, 2010, para. 30) .

Microblogging developed out of a particular blogging practice, wherein bloggers would post small messages or single files on a blog post. Blogs that focused on such “microposts” were then termed “tumblelogs” and were described as “a quick and dirty stream of consciousness” kind of blogging (Kottke, 2005, para. 2)

most popular microblogs are Twitter (launched in 2006), tumblr (launched in 2007), FriendFeed (launched in 2007 and available in several languages), Plurk (launched in 2008 and popular in Taiwan), and Sina Weibo (launched in 2009 and popular in China).

Conference chatter (backchanneling) is another widely studied area in the realm of scholarly microblogging. Twitter use at conferences is generally carried out by a minority of participants

Wikis are collaborative content management platforms enabled by web browsers and embedded markup languages.

Wikipedia has been advocated as a replacement for traditional publishing and peer review models (Xia o & Askin, 2012) and pleas have been made to encourage experts to contribute (Rush & Tracy, 2010) . Despite this, contribution rates remain low — likely hindered by the lack of explicit authorship in Wikipedia, a cornerstone of the traditional academic reward system (Black, 2008; Butler, 2008; Callaway, 2010; Whitworth & Friedman, 2009) . Citations to scholarly documents —another critical component in the reward system —are increasingly being found i n Wikiped ia entries (Bould et al., 2014; Park, 2011; Rousidis et al., 2013) , but are no t yet seen as valid impact indicators (Haustein, Peters, Bar -Ilan, et al., 2014) .

The altmetrics manifesto (Priem et al., 2010, para. 1) , altmetrics can serve as filters , which “reflect the broad, rapid impact of scholarship in this burgeoning ecosystem”.

There are also a host of platforms which are being used informally to discuss and rate scholarly material. Reddit, for example, is a general topic platform where users can submit, discuss and rate online content. Historically, mentions of scientific journals on Reddit have been rare (Thelwall, Haustein, et al., 2013) . However, several new subreddits —e.g., science subreddit 4 , Ask Me Anything sessions 5 –have recently been launched, focusing on the discussion of scientific information. Sites like Amazon (Kousha & Thelwall, 2015) and Goodreads (Zuccala, Verleysen, Cornacchia, & Engels, 2015) , which allow users to comment on and rate books, has also been mined as potential source for the compilation of impact indicators

libraries provide services to support researchers’ use of social media tools and metrics (Lapinski, Piwowar, & Priem, 2013; Rodgers & Barbrow, 2013; Roemer & Borchardt, 2013). One example is Mendeley Institutional Edition, https://www.elsevier.com/solutions/mendeley/Mendeley-Institutional-Edition, which mines Mendeley documents, annotations, and behavior and provides these data to libraries (Galligan & Dyas -Correia, 2013) . Libraries can use them for collection management, in a manner similar to other usage data, such as COUNTER statistics (Galligan & Dyas -Correia, 2013) .

Factors affecting social media use; age, academic rank and status, gender, discipline, country and language,