#Recovery: Recovering in and with Tumblr

Team members: Natalia Sanchez-Querubin, Jelena Schuhmacher, Giacomo Traldi, Carla McCutchin, and Sheena Manabat

Introduction:

Our project addresses the use of Tumblr as platform for recovery. Users within this space have formed communities around the notion of recovery by documenting their own achievements, obstacles, and goals through images and text. On Tumblr, recovery is not only digital but it is also networked. Hashtags serve to narrate images as well as to insert them in discursive and technical streams, to be found by chance or choice in a Tumblr user’s dashboard. Consequently, by using the #recovery Tumblr users are forming communities of healing and support.

Through the efforts of its users Tumblr has come to be known for it’s mental health community. This has been documented by various media outlets as seen below:

Research Questions:

1. How is tumblr used for #recovery processes by its users?

2. Are social media functioning as tools for recovery? If so, what do they enable?

3. How is the recovery outside of the medical context enriched by platform vernacular?

Methodology: Co-hashtag analysis and content analysis for studying recovery on Tumblr.

The project consist of two analytical layers each with their own method. It begins by producing a co-hashtag network graph using the broad query #mentalillness on Tumblr. The most connected nodes in the network included hashtags related to specific conditions such as depression and eating disorders and to terms associated with the umbrella term recovery, including #recovery, #selfcare and #therapy. What followed is the production of a second query now around the term #recovery with the outputs being, first, a co-hashtag network and, second, the capture of the all the images associated with the hashtag along with their metadata.

The #recovery co-hashtag network was edited using Gephi. The force Atlas Algorithm pushed connected nodes to the center and the least connected nodes were filtered. The modularity function was then ran in order to identify communities of strongly connected nodes. Afterwards the each of the five communities were manually analyzed and the hashtags contained in them categorized according to the following criteria:

a) Recovering from..

b) Recovery related terms, such as self-love

c) recovery tools, for example, art and yoga

d) Affirmations — I am a recovery warrior.

These terms were then arranged in a narrative form and used to produced what we have termed, ‘hashtags monologues’. These are co-hashtag networks, ran through a modularity algorithm, read by machines to tell five collective stories about #recovery. To produce the videos the hashtags, now in their narrative form, were inputted in Google’s text to speech tool and used as a voice-over for the screen-captures of the navigation within the hashtag networks in Gephi. The materials are edited on Premier. Ultimately, these five videos are offered as a visualization of what can be interpreted as modes of publishing and tagging content about recovery, and in themselves, perhaps, as indications of how health is discussed, performed as well as cared for in Tumblr.

The second analytical layer attends to the visual components of the network. All images captured were manually classified according to the following questions: how is recovery sustained? Six main themes emerged:

1. Textual post that include inspirational quotes, memes, short reflexions.

2. Images of the body, of which the selfie was the predominant format.

3. Tools or activities for recovering. Knitting, watching television, writing are some of the examples found.

4. Food

5. Art

6. Humor

The following step was to focus on the images that were placed on the category of ‘selfie’ or #recovery selfie. Here the same categorization process was repeated by related now the selfie images with the previous categories. For example, selfies that includes food. Each sub-group is was then analyzed in more detail by tracing images to their original posts and captures relevant quotes. The results are an annotated typology of how recovery is publicly performed by users through the medium of the selfie.

-- NataliaSanchez - 06 Jul 2016