Title: #doomsdayprepper: Analysing the online prepper community on Instagram1
Abstract
Prepping, the activity of preparing for large-scale disasters, has become a global phenomenon and received increasing attention in academic circles recently. Since online platforms play a key role in this movement, the current study analyses images posted on the social media platform, Instagram. Using machine learning models, the main content of images is identified, as well as the gender of people and the sentiment of images. The findings show that, as depicted on Instagram, 1) weapons, supplies and knowledge-acquisition are important facets to prepping; 2) this is not a male-dominated activity, but rather equally male and female; 3) prepping posts are mostly positive, rather than negative as expected (because prepping involves preparing for worse-case scenarios). Suggestions are also made for further research.
Keywords: prepping, Doomsday Preppers, survivalism, machine learning, big data, social media, online communities, Instagram
Ensovoort, volume 40 (2019), number 11: 1
Introduction
Across the world, an increasing number of people are engaged in prepping – the act of preparing for disasters. According to Fish (2017), 4-9 million Americans are engaged in prepping. The phenomenon is not, however, confined to the US, but extends to countries such as South Korea (Fish, 2017, Neuman, 2018), Ireland (Brophy, 2019), the United Kingdom (Wollaston, 2019, John, 2019), France (Anonymous, 2018), Sweden (Orange, 2018, Ehnroth and Rössner, 2018), Finland (Parkkinen, 2018), Australia (Szabo, 2018, Deutrom, 2018), the Netherlands (Van Dongen, 2018) and South Africa (Suidlanders, 2016). Prepping has moved across the political divide as well: while traditionally associated with the right wing, the left wing is now also part of the prepping movement (Riederer, 2018, Deller, 2019, Sedacca, 2017). In addition, billionaires have recently purchased land in New Zealand and built underground bunkers there (Carville, 2018, Sedacca, 2017). The prepper community has even developed its own terminology, such as BOB (Bug Out Bag), SHTF (Shit Hit The Fan), GOOD (Get Out Of Dodge), and TEOTWAWKI (The End Of The World As We Know It) (Brown, 2017, Parkkinen, 2018).
Following the growth in the popularity of prepping, this phenomenon has also received increasing attention from academics recently, with a number of dissertations (Rogers, 2015, Ehnroth and Rössner, 2018, Parkkinen, 2018), theses (Gonowon, 2011, Imel-Hartford, 2013, Aldousari, 2015, Sims, 2017, Mills, 2017), books (Foster, 2014, Christian, 2016, Joyce, 2018) and articles (Simut, 2012, Senekal, 2014, Aldousari, 2014, Foster, 2016, Kelly, 2016, Mills, 2018, Roberts and Hogan, 2019) published on the subject. In 2015, Emergent BioSolutions also compiled a multi-phase research report on prepping (Fish, 2017).
An online community can be defined as,
… a group of people interested in a particular topic, or that share some ways of thinking, or that in general have some kind of link that brings them together, with the peculiarity that they interface and connect to each other through a data communication network (such as Internet). (Fornacciari, Mordonini, and Tomaiuolo, 2015:53)
The prepper community is an online community that exists through YouTube channels (Sims, 2017:46, Fish, 2017), Facebook (Fish, 2017, Ehnroth and Rössner, 2018), Pinterest (Shade, 2014), online discussion forums (Parkkinen, 2018) and Instagram (Acker and Carter, 2018). Fish (2017) adds that preppers are heavily engaged online and also rely on word of mouth within their community, which makes online social networks particularly important in this movement.
The current study uses machine learning to analyse posts on Instagram tagged with the hashtag #doomsdaypreppers. Using various Convolutional Neural Network (CNN) models, the dominant content, people’s gender, and the sentiment of images are identified.
The article is structured as follows: after providing a background to prepping, including a definition and description of prepping, an overview of data collection methods and machine learning for image recognition is provided. The results and discussion is divided into three themes: content, gender and sentiment, with each providing an overview of this theme in the literature and a discussion of applicable machine learning models and findings from the current study. The article concludes with final remarks and future avenues of research.
Background
The current prepping movement has its origins in Cold War survivalism, where people built bomb shelters and stockpiled supplies in anticipation of a global nuclear war (Deller, 2019). Prepping and survivalism are similar concepts, but Sims (2017:3) notes that preppers wish to distinguish themselves from survivalists to avoid the stigma attached to the latter term, since survivalism has been linked to political extremists and militia members like Timothy McVeigh, David Koresh and Ted Kaczynski (see also Morris, 2012, Riederer, 2018 and Sedacca, 2017). Sims (2017:20) describes survivalism as “a specific reaction or response to a perceived external threat, whether it be impending ecological collapse or a rise in social disorder.” The difference between survivalism and prepping, in Sims’s view, lies in survivalists’ wish to withdraw from society, whereas preppers “identified as preppers as a way to claim an identity that was tied to self-sufficiency and self-reliance within the context of the generalized culture” (2017:4). This emphasis on withdrawing from society can be seen in Peterson’s (1984:44) definition of survivalism: “Survivalists are those who believe that the United States is on the verge of collapse. They hope to survive economic, social, or nuclear disaster by making preparations for self-sufficiency and breaking away as much as possible from mainstream society.” In contrast, preppers emphasise that they remain normal members of society (Ehnroth and Rössner, 2018), and justify their preparations as something sensible to do, akin to an insurance policy (Morris, 2012) (Szabo, 2018) (Fish, 2017) (Xavier, 2017) (Aaron, 2018). Fish (2017) notes: “your average prepper is an ordinary person trying to do his/her best for his/her family by preparing for emergency events” (see also Morris (2012)). As Phil Burns, owner of the American Prepper’s Network (Morris, 2012), argues,
Seriously, why do people have homeowners insurance? It’s so that if something catastrophic happens to your house you can get money to buy a new one – and not be homeless. Prepping is basically the same thing – we educate ourselves and purchase items that will be essential to continue our way of life in a catastrophic event.
Velarde (2013) defines prepping simply: “Preppers, then, are ready or are in the process of making themselves ready (preparing).” More comprehensive is Rahm’s (2013) definition, who defines preppers as, “people who believe in abrupt, imposing and near-in-time disasters and who are actively and practically preparing to survive this imminent apocalypse.” These preparations include for large-scale disasters such as a global economic collapse, nuclear war, major terrorist attacks (often using biological weapons), worldwide pandemics, solar flares or the eruption of super volcanoes. Some preppers also prepare for more unlikely scenarios such as alien attacks or a zombie apocalypse, although these appear to be a minority in the prepper community. As Velarde (2013) and Mills (2018) recognise, however, preparations do not necessarily have to be for such global disasters and include preparations for more localized crises such as tsunamis, earthquakes and hurricanes. Velarde (2013) concludes, “Not all preppers, then, are readying for doomsday per se, but they are nevertheless preppers in the sense that they are actively preparing to survive some kind of disaster.” An example of a preparation for a smaller emergency is the preparations made for Brexit by UK preppers, where preppers fear a disruption in the supply line (Wollaston, 2019, John, 2019).
Although only a fringe minority within the prepping community prepares for a zombie outbreak, the phenomenon of a zombie outbreak has become a metaphor for disaster preparedness in general (Rodriguez, 2014, Sims, 2017:199). For instance, the New York Zombie Outbreak Response Team (ZORT) was founded in May 2013 and “uses the zombie apocalypse as a metaphor for everything from attackers to natural disasters or terrorist attacks” (Rodriguez, 2014). Stickers and patches indicating “Zombie Outbreak Response Team” are also widely available on prepper websites. Nor is this metaphor confined to civilians: The US military’s CONPLAN 8888-11, which details emergency procedures in the event of a zombie outbreak, was developed as a fictitious scenario to develop scenario planning (Carter, 2018, Crawford, 2014, and Lobold, 2014). The Centre for Disease Control and Prevention (CDC) used preparedness for a zombie outbreak to market preparedness (Kruvand and Silver, 2013, Kruvand and Bryant, 2015), arguing,
There are all kinds of emergencies out there that we can prepare for. Take a zombie apocalypse for example. That’s right, I said z-o-m-b-i-e a-p-o-c-a-l-y-p-s-e. You may laugh now, but when it happens you’ll be happy you read this, and hey, maybe you’ll even learn a thing or two about how to prepare for a real emergency (Centre for Disease Control and Prevention, 2011).
The Federal Emergency Management Agency (FEMA) concurs with the CDC and also encourages people to prepare for a zombie attack, noting, “Preparing for a zombie attack, or other fictional disasters, can provide useful tips to get prepared for a real disaster” (Federal Emergency Management Agency, 2011).
Prepping has become a major business and companies have been founded that cater specifically to the prepping community. According to Timmer (2018), companies that cater to preppers have grown their revenue by about 700% over the last decade. US Generator manufacturer Generac Holdings reported annual sales growth of over 30% in 2012, while Hardened Structures claimed an increase of 15% in the same year (Taylor, 2012). Fish (2017) found that preppers spend significant amounts of money in preparing for disasters:
- Fifty percent of preppers in their study spent more than $500 per year on supplies.
- 15%-30% of preppers spent more than $1,000 annually on supplies; with an average spend of $1,850.
- Outliers spend vastly more. In Fish’s research, one man spent $10,000 annually on supplies, and another one spent $75,000.
Prepping gear constitutes an important part of the activities of the prepping community and online forums often compare, review and discuss various tools and other items. Brown (2017) notes that Cold War survivalists were some of the first to adopt new technologies for this purpose,
The communities that formed around the survivalist newsletters of the 1970s were early adopters of BBS and Usenet, precursors to internet forums. The most well known early online community was the WELL, which stood for Whole Earth ’Lectronic Link. It was associated with the Whole Earth Catalog, a magazine focused on self-sufficiency, survivalism, and sustainability. Today, there are hundreds of active prepper forums, like Zombie Squad, Survivalist Boards, and American Prepper Network, where users share tips, discuss Armageddon scenarios, and recommend supplies. But now the most active conversations happen in Facebook groups and Reddit threads.
Because equipment and skills are such important components of prepping, and because online forums constitute a key contact environment for the prepping community, an analysis of social media will shed light on prepping practice. The following section discusses how this analysis was done.
Methods
Instagram was founded in 2010 by Kevin Systrom and Mike Krieger and gathered over a million members in less than three months (Latiff and Safiee, 2015:14). By 2016, Instagram already had 400 million active users worldwide, compared with Twitter’s 320 million at that time (Phua, et al., 2017:412). This platform allows users to share photos and short videos with followers, comment, or just indicate that they like the post. About 40 million photos are currently posted to Instagram daily.
Because Instagram is such a young platform, research on the value of Instagram is still very limited. The majority of Instagram studies focus on identity creation (Fallon, 2014), socialising (Lup, et al., 2015), and marketing (Bergström and Bäckman, 2013, Latiff and Safiee, 2015, Phua, et al., 2017). The current study ties in with identity creation and socialising, since Instagram is used by the prepping community to engage with other like-minded individuals. Sims (2017:45) notes,
“Interacting with others who had similar interests was a key way that these preppers developed their skillset and knowledge base.”
Instagram is especially a platform used by young people, and teenagers in particular prefer Instagram to Facebook (Duncan, 2016). Research by Alhabash and Ma (2017) has shown that Instagram is used especially for entertainment, self-expression, social interaction, and the sharing of information. Research by Hu, Manikonda and Kambhampati (2014) has shown that mainly eight different kinds of photos are placed on Instagram: self-portraits, friends, activities, pictures with built-in text, food, fashion and pets. The most popular type of photo is the selfie (Hu, et al., 2014).
Instagram posts with the hashtag #doomsdayprepper were downloaded on 29 March 2019, with a total of 9 020 posts. This hashtag was chosen because the reality television show, Doomsday Preppers, which was aired on the National Geographic Channel from 2011 to 2014, occupies a key position in the literature on the prepper movement. This show is often referenced by scholars and preppers, e.g. Simut (2012), Foster (2014), Aldousari (2015), Kelly (2016), Foster (2016), Sims (2017), Mills (2018), Acker and Carter (2018), and Joyce (2018). Sims (2017:217) for instance notes that this show inspired her study, while Aldousari (2015), Kelly (2016) and Christian (2016) devote their entire studies to this show. Choosing this hashtag allowed for the collection of all Instagram posts that are deliberately placed in the prepper discourse.
Figure 1 shows the number of posts over time, as well as the type of posts. The first post was made on 2012-02-09 and the last on 2019-03-29. Usernames and biographical information were not collected for ethical reasons.
Figure 1 shows that the number of posts steadily increased over time (only 3 months are taken into account for 2019). This may be due to Instagram’s increasing popularity, or it may indicate an increased interest in the subject. The bottom graph shows the number of posts per day, and here it can be seen that no more than 30 posts are made with this hashtag per day. The majority of posts are in image format, with only 4,39% of posts relating to videos. Video has become slightly more popular over time, but the dataset remains dominated by images for the whole period. The rest of this article focuses on images.
After collecting posts, images were classified using machine learning technology. Machine learning is a subfield of artificial intelligence (AI) and was developed from the 1960s (Michie, 1968, and Kononenko 2001), in particular through the works of Rosenblatt (1962), Nilsson (1965), and Hunt, Martin, and Stone (1966). The field gained ground in the last two decades because of the big data revolution (Jordan and Mitchell, 2015:256), leading Jordan and Mitchell (2015:260) to claim: “machine learning is likely to be one of the most transformative technologies of the 21st century”.
Machine learning models require large amounts of training data to develop, in particular where images are concerned (Beam and Kohane, 2018). Since the current objective is to study the prepping community rather than to develop new machine learning models, I use existing models developed and tested in the Computer Science community.
Results and discussion
Gear
The most commonly mentioned practice identified in Sims’s study, was the collection and storage of an emergency ration of food and water: “There were variations throughout the sample related to the importance of various other practices, but ensuring that you have enough food and water for all of the people in your household was the foundation upon which all other prepping practices were built” (Sims, 2017:28). The second most important activity of prepping in Sims’s study, was defence. In particular, guns “played a primary role in the culture of prepping” (Sims, 2017:40). Sims also found acquiring knowledge to be a key component of prepping, “Acquiring knowledge was a key process of prepping and merely having the material objects related to prepping is not enough to ensure one’s survival” (Sims, 2017:45).
Sims’s findings show the importance of having the right equipment, and showcasing equipment and preps is one of the most common themes in online discussion forums (Imel-Hartford, 2013:78, and Parkkinen, 2018:31). In line with Sims (2017:40-43), Roberts and Hogan (2019:3) also note the importance of weapons,
A fondness for weaponry of all kinds and means of self-defense are often at the center of the preparations and infrastructure, so that the prepared may defend themselves not only against an enemy, but also against those who were not so well prepared for calamity and unwisely attempt to seek material support or other assistance from their fellows.
In addition, Brown (2017) reports an obsession with clothing in the prepper community and quotes one prepper as saying,
In hunting there’s a fashion-show element, because everyone’s got to have cooler camo than the other guy — the latest computer-generated camo. Backpackers are more obsessed with absolute performance and they don’t care what it looks like. They can wear some pretty goofy-looking stuff. But preppers tend to focus on the appearance of clothing because they’re more aware of what their clothes signal about them, and they’re trying to manage that signal.
Because equipment, stockpiles and weapons are so important, the content of Instagram posts was first determined. I used the deep Convolutional Neural Networks (CNN) model developed using the MobileNet architecture described by Howard, Zhu, Chen, Kalenichenko, Wang, Weyand, Andreetto and Adam (2017). This model achieved a 70,6% accuracy in classifying the content of images, which is higher than for most of the other models it was compared with. I only use the first classification level, however: if a picture for instance contains a knife, stove and tent, I use the object that the model identified as most prominent.
Table 1 provides the top 40 categories identified. The description refers to the label assigned by the model. In total, 639 categories were labelled by the model.
Table 1 The top 40 items identified
|
Description |
#records |
%records |
|
412 |
4.78% |
|
335 |
3.88% |
|
315 |
3.65% |
|
260 |
3.01% |
|
206 |
2.39% |
|
191 |
2.21% |
|
175 |
2.03% |
|
169 |
1.96% |
|
135 |
1.57% |
|
122 |
1.41% |
|
109 |
1.26% |
|
90 |
1.04% |
|
85 |
0.99% |
|
85 |
0.99% |
|
74 |
0.86% |
|
66 |
0.77% |
|
65 |
0.75% |
|
64 |
0.74% |
|
62 |
0.72% |
|
61 |
0.71% |
|
58 |
0.67% |
|
56 |
0.65% |
|
46 |
0.53% |
|
44 |
0.51% |
|
43 |
0.50% |
|
43 |
0.50% |
|
41 |
0.48% |
|
41 |
0.48% |
|
40 |
0.46% |
|
40 |
0.46% |
|
39 |
0.45% |
|
38 |
0.44% |
|
38 |
0.44% |
|
38 |
0.44% |
|
36 |
0.42% |
|
35 |
0.41% |
|
35 |
0.41% |
|
33 |
0.38% |
|
32 |
0.37% |
|
32 |
0.37% |
The objects identified support previous qualitative studies by showing similar items of importance. In line with Sims’s (2017:45) finding that acquiring knowledge is one of the most important facets of prepping, web sites (1) and books (2) are the most prominent. As can be expected, guns feature high on the list (3, 4, 12, 13), along with bullet proof vests (20). Other supplies such as food (34), medicine (28, 32, 33), and gas masks (9) are also prominent. Bug out bags (11) and bug out vehicles (8) also feature high on the list. In addition, military uniforms (15) are also prominent, showing Brown’s (2017) assertion that clothing plays an important role: preppers often dress in tactical or camouflage uniforms.
The content of Instagram posts shows the primacy of defence, knowledge-acquisition, supplies and equipment.
Masculinity
Numerous scholars note the importance of masculinity in the prepping community. Sims (2017:14) refers to prepping as “this seemingly male dominated and marginalized culture,” although she (2017:120) finds that this is not the case. Kelly (2016:98) argues, “Prepper discourse encourages the development of masculine-coded abilities, including mechanical labor, wilderness training, and weapons proficiency.” She (2016:107) describes the television show, Doomsday Preppers,
In summary, these episodes begins (sic) with a beleaguered father/husband’s confession of his apocalyptic anxiety, followed by scenes of everyday life where participants move fluidly between masculine archetypes (father, husband, laborer, soldier, and priest), and conclude with a series of staged preparedness rituals. These performances enact the fears expressed in the one-on-one interviews, providing audiences with examples of how to translate apocalyptic fears into productive models of self-made manhood.
Because gender plays such an important role in the prepping community, I determined the gender of people in images. Levi and Hassner (2015) developed a model with which to classify people’s age and gender by also making use of CNN. Their model was trained on the Adience benchmark for age and gender classification of face images (Eidinger, et al., 2014), comprising approximately 26 000 images of 2 284 human subjects. Their results show an accuracy level of 86,8% for gender estimation and 50,7% accuracy for age estimation, which increases to 84,7% if the model is allowed to be one age group off. A survey of age classification using machine learning can be found in Han, Otto and Jain (2013), while a survey of gender classification research using machine learning can be found in Reid, Samangooei, Chen, Nixon and Ross (2013).
Of the 9 020 posts tagged with #doomsdayprepper, only 2 308 contained images of people. Of these, 1 178 (51,04%) were identified as male, and 1 130 (48,96%) as female. Figure 2 shows the results of the gender model, with examples. To ensure privacy, examples were chosen that do not show faces clearly.
Figure 2 shows that the number of male subjects in these posts increased sharply in 2018, but in general, the gender distribution of people in these images is roughly 50/50. The examples in Figure 2 below show typical prepping activities and scenarios. Note the male image with the vehicle: this is a screen capture from the movie Mad Max Fury Road (2015), a typical post-apocalyptic movie. The female examples show the same activities and scenarios as the male images: using firearms, gas masks and post-apocalyptic scenarios. Incidentally, the image of a woman firing a pistol in an indoor shooting range is the first post with the hashtag #doomsdayprepper.
Females are clearly incorporated in this discourse, showing that while the activities are masculine, the subjects are female in about half of the images. The results therefore show that this discourse is not male-dominated.
Sentiment
One might expect images uploaded by preppers to be negative, since prepping ultimately involves thinking about worse-case scenarios. Aldousari (2015:30) writes,
Thinking of doomsday causes preppers to reflect on the meanings and value of life. These reflections are usually accompanied by negative emotions such as fear and anger.
Aldousari (2015:17) continues,
The prepping journey consumed every aspect of preppers’ life: From the moment they embarked on the prepping journey, they were continuously stockpiling food, learning new survival skills, and sacrificing much of life’s pleasures in order to be ready for the impending disaster. This lifestyle caused anxiety and stress that many researchers have shown to influence consumption behaviour.
Not surprisingly, Kruvand and Bryant (2015) found that preppers score high on anxiety. Because prepping is associated with negative emotions, I wanted to determine whether negativity could be identified in pictures.
Using the model developed by Campos, Jou and Giró-i-Nieto (2017), I classify images according to sentiment. Campos, Jou and Giró-i-Nieto’s (2017) trained a CNN-model on 1 269 images, annotated using Amazon Mechanical Turk annotators, in order to classify images as either positive or negative. Determining whether images are positive or negative involves a judgement on the part of the annotator, and hence one can expect diverging results. To compensate for the inherent subjectivity of this task, the authors tested their model on images where five annotators agreed on the sentiment of an image, where only four annotators agreed, and where only three annotators agreed. Their model shows a correlation between human and model of 0,83 where five annotators agreed, 0,78 where four annotators agreed, and 0,75 where four annotators agreed. This means that the model is highly accurate in determining the sentiment of images, but accuracy is reduced on more ambivalent images where human annotators also could not agree.
Figure 3 shows the results of the sentiment model, with a few examples.
Almost 70% of images were classified as positive. The examples on the bottom left show preparations and family, which one could argue are positive because prepping empowers a person to care for loved ones in uncertain times. Senekal (2014:193) concludes after studying preppers in the run up to the 1994 election in South Africa, “This was a very uncertain time in South Africa’s history and perhaps preparations provided comfort in the face of this uncertainty.” Similarly, Aldousari (2015:21) notes that “prepping enhances preppers’ anxiety buffer,” and Sims (2017:266) concurs with this view, arguing that preppers “focused on survival as a way to assuage their existential anxieties.” This tendency to assuage fears by actively preparing themselves for difficult times could be the reason why images are predominantly positive. The negative example images on the right, however, show post-apocalyptic scenes and men dressed for a post-apocalyptic world, which is a decidedly negative environment. Although these images do occur frequently, they make up less than a third of Instagram posts.
Conclusion
This study analysed Instagram posts with the hashtag #doomsdayprepper. It was shown that the content of posts relate to what previous qualitative researchers have found, in particular in showing the importance of weapons, supplies, clothing and knowledge acquisition. In addition, it was shown that the subjects in images are evenly distributed between male and female, showing that females have been incorporated in this discourse and that it is not a male-dominated environment. Thirdly, by identifying the sentiment in images, it was shown that images are predominantly positive, indicating that this community is less focused on the negative aspects of prepping and more on the positive.
The current study only uses one hashtag, however, whereas there are numerous hashtags associated with the prepper movement. Future research could incorporate more hashtags and larger datasets, which might lead to more nuanced results. In addition, other online communities could be studied in a similar manner.
Endnote
1Ethical clearance was obtained for this study with number UFS-HSD2019/0175.
Bibliography
Aaron, P., 2018. Being a prepper is like buying insurance. [Online]
Available at: https://www.thesimpleprepper.com/prepper/essentials/being-a-prepper-is-like-buying-insurance/ [Accessed 20 March 2019].
Acker, A. and Carter, D., 2018. Pocket Preppers: Performing preparedness with everyday carry posts on Instagram. Proceedings of the 9th International Conference on Social Media and Society,:207-211.
Aldousari, A. A., 2014. Exploring Consumptive behavior under perceived threat: the case of Doomsday Preppers. Advances in Consumer Research, 42:769-769.
Aldousari, A. A., 2015. Exploring consumer behavior under perceived threat. Unpublished PhD-thesis: Rutgers University.
Alhabash, S. and Ma, M., 2017. A tale of four platforms: motivations and uses of Facebook, Twitter, Instagram, and Snapchat among college students. Social Media + Society,:1-13.
Anonymous, 2018. Fearing worst, French ‘preppers’ gear up for the Day After. [Online]
Available at: https://www.france24.com/en/20180323-fearing-worst-french-preppers-gear-day-after [Accessed 19 March 2019].
Beam, A. L. and Kohane, I. S., 2018. Big Data and machine learning in health care. Journal of the American Medical Association, 319(13):1317-1318.
Bergström, T. and Bäckman, L., 2013. Marketing and PR in Social Media. How the utilization of Instagram builds and maintains customer relationships. Unpublished BA-dissertation: Stockholms Universitet.
Brophy, D., 2019. ‘We’re really only nine meals removed from total anarchy’: Ireland’s preppers and survivalists. [Online]
Available at: https://www.thejournal.ie/irish-preppers-and-survivalists-brexit-weather-4443870-Jan2019/ [Accessed 19 March 2019].
Brown, J., 2017. The prepper obsession with clothes. [Online]
Available at: https://www.racked.com/2017/5/29/15663024/prepper-clothes [Accessed 19 March 2019].
Campos, V., Jou, B. and Giró-i-Nieto, X., 2017. From pixels to sentiment: Fine-tuning CNNs for visual sentiment prediction. Image and Vision Computing, Volume 65,:15-22.
Carter, S. L., 2018. Are we prepared for the zombie apocalypse? Actually, yes. [Online]
Available at: https://www.bloomberg.com/opinion/articles/2018-11-17/zombie-apocalypse-yes-the-u-s-is-prepared [Accessed 20 March 2019].
Carville, O., 2018. The super rich of Silicon Valley have a doomsday escape plan. [Online]
Available at: https://www.bloomberg.com/features/2018-rich-new-zealand-doomsday-preppers/ [Accessed 19 March 2019].
Centre for Disease Control and Prevention, 2011. Preparedness 101: Zombie Apocalypse. [Online]
Available at: https://blogs.cdc.gov/publichealthmatters/2011/05/preparedness-101-zombie-apocalypse/ [Accessed 20 March 2019].
Christian, T. A., 2016. The recuperation of hegemonic masculinity on Doomsday Preppers. In: L. A. Clark, A. Firestone and M. F. Pharr, eds. The Last Midnight: Essays on Apocalyptic Narratives in Millennial Media. Jefferson: McFarland and Co:48-59.
Crawford, J., 2014. Pentagon document lays out battle plan against zombies. [Online]
Available at: https://edition.cnn.com/2014/05/16/politics/pentagon-zombie-apocalypse/index.html [Accessed 20 March 2019].
Deller, A. S., 2019. Doomsday Preppers of the Trump Presidency. [Online]
Available at: https://medium.com/@tdeller/doomsday-preppers-of-the-trump-presidency-877ff56287e9 [Accessed 19 March 2019].
Deutrom, R., 2018. Aussie preppers show us their kits. [Online]
Available at: https://www.news.com.au/technology/innovation/aussie-preppers-show-us-their-kits/news-story/b3f6ab1bf2eb9a4ca0b2ed02a0370e1e [Accessed 19 March 2019].
Duncan, F., 2016. So long social media: The kids are opting out of the online public sphere. [Online]
Available at: http://theconversation.com/so-long-social-media-the-kids-are-opting-out-of-the-online-public-square-53274 [Accessed 17 September 2018].
Ehnroth, M. and Rössner, C. J., 2018. ”Jag är en normal människa” Drivkrafter och motivation för svenska preppers. Unpublished dissertation: Stockholms universiteit.
Eidinger, E., Enbar, R. and Hassner, T., 2014. Age and gender estimation of unfiltered faces. Transactions on Information Forensics and Security, 9(12).
Fallon, K., 2014. Streams of the Self: the Instagram feed as narrative autobiography. Toronto, University of Toronto:54–60.
Federal Emergency Management Agency, 2011. From CDC – Preparedness 101: Zombie Apocalypse. [Online]
Available at: https://www.fema.gov/blog/2011-05-19/cdc-preparedness-101-zombie-apocalypse [Accessed 20 March 2019].
Fish, R., 2017. Preppers: On the frontline of U.S. preparedness. [Online]
Available at: https://globalbiodefense.com/2017/10/27/preppers-on-the-frontline-of-u-s-preparedness/ [Accessed 19 March 2019].
Fornacciari, P., Mordonini, M. and Tomaiuolo, M., 2015. A Case-Study for Sentiment Analysis on Twitter. Naples, s.n.:53-58.
Foster, G. A., 2014. Hoarders, Doomsday Preppers, and the culture of apocalypse. New York: Palgrave MacMillan.
Foster, G. A., 2016. Consuming the apocalypse, marketing bunker materiality. Quarterly Review of Film and Video, 33(4):285-302.
Gonowon, C., 2011. Just in Case: An empirical-phenomenological analysis of the lived experience of modern survivalists. Unpublished PhD-thesis: Chicago School of Professional Psychology.
Han, H., Otto, C. and Jain, A. K., 2013. Age estimation from face images: Human vs. machine performance. s.l., 2013 International Conference on Biometrics (ICB).
Howard, A. G. et al., 2017. MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications. arXiv Preprints.
Hu, Y., Manikonda, L. and Kambhampati, S., 2014. What we Instagram: A first analysis of Instagram photo content and user types. Ann Arbor, The AAAI Press,:595-598.
Hunt, E., Martin, J. and Stone, P., 1966. Experiments in Induction. New York: Academic Press.
Imel-Hartford, L., 2013. The Preppers: A multiple case study of individuals who choose a moderate survivalist lifestyle. Unpublished PhD-thesis: Northcentral University.
John, T., 2019. Anxious Brits have bought hundreds of food-prepper ‘Brexit Boxes’. [Online]
Available at: https://edition.cnn.com/2019/01/14/uk/brexit-box-stockpile-preppers-gbr-intl-scli/index.html [Accessed 19 March 2019].
Jordan, M. I. and Mitchell, T. M., 2015. Machine learning: Trends, perspectives, and prospects. Science, 349(6245):255-260.
Joyce, S., 2018. Transmedia storytelling and the apocalypse. London: Palgrave Macmillan.
Kelly, C. R., 2016. The man-pocalpyse: Doomsday Preppers and the rituals of apocalyptic manhood. Text and Performance Quarterly, 36(2):95-114.
Kononenko, I., 2001. Machine learning for medical diagnosis: history, state of the art and perspective. Artificial Intelligence in Medicine, 23(1):89-109 .
Kruvand, M. and Bryant, F. B., 2015. Zombie Apocalypse: Can the undead teach the living how to survive an emergency?. Public Health Report, 120(6):655–663.
Kruvand, M. and Silver, M., 2013. Zombies gone viral: How a fictional zombie invasion helped CDC promote emergency preparedness. Case Studies in Strategic Communication, Volume 2:34-60.
Latiff, Z. A. and Safiee, N. A. S., 2015. New business set up for branding strategies on social media – Instagram. Procedia Computer Science, Volume 72:13-23.
Levi, G. and Hassner, T., 2015. Age and gender classification using Convolutional Neural Networks. Boston, s.n.
Lobold, G., 2014. Exclusive: The Pentagon has a plan to stop the zombie apocalypse. Seriously.. [Online]
Available at: https://foreignpolicy.com/2014/05/13/exclusive-the-pentagon-has-a-plan-to-stop-the-zombie-apocalypse-seriously/ [Accessed 20 March 2019].
Lup, K., Trub, L. and Rosenthal, L., 2015. Instagram #Instasad?: Exploring associations among Instagram use, depressive symptoms, negative social comparison, and strangers followed. Cyberpsychology, Behavior and Social Networking, 18(5),:247-252.
Michie, D., 1968. “Memo” functions and machine learning. Nature, 218(5136):19-22.
Mills, M. F., 2017. Witness to the American apocalypse? : a study of 21st century ‘doomsday’ prepping. Unpublished PhD-thesis: University of Kent.
Mills, M. F., 2018. Preparing for the unknown… unknowns: ‘doomsday’ prepping and disaster risk anxiety in the United States. Journal of Risk Research.
Morris, C., 2012. Businesses that cater to doomsdayers. [Online]
Available at: https://www.cnbc.com/id/48471717 [Accessed 19 March 2019].
Neuman, M., 2018. Seoul’s ‘War Preppers’ are still (sort of) expecting the worst. [Online] Available at: https://www.citylab.com/life/2018/06/seouls-war-preppers-are-still-sort-of-expecting-the-worst/562763/ [Accessed 19 March 2019].
Nilsson, N., 1965. Learning Machines. New York: McGraw-Hill.
Orange, R., 2018. Sweden’s preppers stock up for potential Russian attack. [Online]
Available at: https://www.telegraph.co.uk/news/2018/06/18/swedens-preppers-stock-potential-russian-attack/ [Accessed 19 March 2019].
Parkkinen, M., 2018. Encountering multiple futures. Finnish survivalism as an anticipatory approach. Unpublished MA-dissertation: University of Turku.
Peterson, R. G., 1984. Preparing for apocalypse: Survivalist strategies. Free Inquiry in Creative Sociology, May, 12(1):44-46.
Phua, J., Jin, S. V. and Kim, J., 2017. Gratifications of using Facebook, Twitter, Instagram, or Snapchat to follow brands: The moderating effect of social comparison, trust, tie strength, and network homophily on brand identification, brand engagement, brand commitment, and membership intentio. Telematics and Informatics, 34:412-424.
Rahm, L., 2013. Who will survive? On bodies and boundaries after the apocalypse. Gender Forum, 45.
Reid, D. et al., 2013. Soft biometrics for surveillance: an overview. In: Machine learning: theory and applications. s.l.:Elsevier:327–352.
Riederer, R., 2018. Doomsday goes mainstream. [Online]
Available at: https://www.dissentmagazine.org/article/doomsday-goes-mainstream-liberal-preppers-silicon-valley [Accessed 19 March 2019].
Roberts, S. T. and Hogan, M., 2019. Left behind: Futurist fetishists, prepping and the abandonment of earth. b2o, 4(2).
Rodriguez, A., 2014. Hurricanes! Terror attacks! Zombies!. [Online]
Available at: http://boroughbuzz.nycitynewsservice.com/2014/04/11/hurricanes-terror-attacks-zombies/ [Accessed 20 March 2019].
Rogers, E. B., 2015. To prepare or not to prepare? A dramatistic rhetorical analysis of National Geographic’s “Doomsday Preppers”. Unpublished MA-dissertation: University of South Alabama.
Rosenblatt, F., 1962. Principles of Neurodynamics. Washington: Spartan Books.
Sedacca, M., 2017. The new doomsayers taking up arms and preparing for catastrophe: American liberals. [Online]
Available at: https://qz.com/973095/the-new-doomsayers-taking-up-arms-and-preparing-for-catastrophe-american-liberals/ [Accessed 19 March 2019].
Senekal, B. A., 2014. An ark without a flood : white South Africans’ preparations for the end of white-ruled South Africa. Journal for Contemporary History, 34(2),:177 – 195.
Shade, C., 2014. Survivalists are using Pinterest to prepare for the apocalypse. [Online] Available at: https://www.theatlantic.com/technology/archive/2014/05/prepping-for-the-apocalypse-on-pinterest/370937/ [Accessed 19 March 2019].
Sims, A. A., 2017. Survival of the Preppers: An exploration into the culture of prepping. Unpublished PhD-thesis: University of Missouri.
Simut, A., 2012. Psycho-apocalypses, from theory to practice. Ekphrasis. Images, Cinema, Theory, Media, Issue 2:64-75.
Suidlanders, 2016. Wie is ons?. [Online]
Available at: https://www.suidlanders.co.za/wie-is-ons/ [Accessed 28 February 2019].
Szabo, S., 2018. Endgame: how Australian preppers are bugging out and hunkering down. [Online]
Available at: https://www.theguardian.com/society/2018/oct/29/end-game-how-australian-preppers-are-bugging-out-and-hunkering-down [Accessed 19 March 2019].
Taylor, C., 2012. Hurricane Sandy Doesn’t Worry Safety-Conscious ‘Preppers’ Homeowners. [Online]
Available at: https://www.insurancejournal.com/news/national/2012/10/28/268263.htm [Accessed 19 March 2019].
Timmer, J., 2018. What exactly are “preppers” prepping for?. [Online]
Available at: https://arstechnica.com/science/2018/05/what-exactly-are-preppers-prepping-for/ [Accessed 19 March 2019].
Van Dongen, A., 2018. Een noodsituatie? Deze ‘doomsday prepper’ is wél overal op voorbereid. [Online]
Available at: https://www.ad.nl/binnenland/een-noodsituatie-deze-doomsday-prepper-is-wel-overal-op-voorbereid~a3de3a59/ [Accessed 19 March 2019].
Velarde, R., 2013. Preparing for the apocalypse: A look at the rise of Doomsday Preppers. Christian Research Journal, 26(4).
Wollaston, S., 2019. ‘I don’t trust the government to look after me or my dog’: meet the Brexit stockpilers. [Online]
Available at: https://www.theguardian.com/politics/2019/jan/14/i-dont-trust-the-government-to-look-after-me-or-my-dog-meet-the-brexit-stockpilers [Accessed 19 March 2019].
Xavier, E., 2017. Preppers: What types of insurance do you have?. [Online]
Available at: https://morethanjustsurviving.com/prepper-question-insurance-types/ [Accessed 19 March 2019].