Exhausting a Crowd

RESEARCH METHODS 

Data Scraping 
Qualitative Content Analysis

MEMBERS 

Nicola Law: Researcher

TIMEFRAME

Nov 8, 2016 - Dec 13, 2016

SUMMARY

With the data scraping method, this research unfolded the privacy invasion concerns in online crowdsourced art, Exhausting a Crowd. It allows users to annotate footage from street surveillance cameras around the world anonymously. The findings revealed that users freely expressed their concerns and critiques related to privacy and surveillance issues politely and sarcastically through a grounded theory analysis.

The javascript code is attached in this study to allows researchers who are interested in pursuing further. Other privacy relevant project was mentioned at the end.

Project Detail Description 

People often take technologies for granted, not being aware of the consequence of the use of technologies. The perspective of privacy calculus suggests that people will disclose private information based on an analysis of benefits and risks (Culnan & Armstrong, 1999). However, how can we do the calculus if we do not know the advantages and risks? How can we see if we can decide what kinds of information to share? Are we still in control of our privacy? Exhausting a Crowd, an online crowdsourced art, reveals every action of people under public surveillance, which illustrates the idea of the nontransparent information interchange in cyberspace - you might not be the information sender. At the same time, your information also becomes transparent for those who has access or the technological skills to archive data.

Experimental Crowdsourced Art – Exhausting a Crowd

Kyle McDonald’s Exhausting a Crowd is inspired by George Perec’s experimental literature, An Attempt at Exhausting a Place in Paris (1974). This 60-pages essay documents observational details that go unnoticed typically when he sat in Saint-Sulpice Square in Paris –buses come and go, tourists are taking photographs, people passing by, empty benches, etc.

Exhausting a Crowd is an online crowdsourced art that allows users to annotate anything in the pre-recorded 12 hours’ footage recorded by the street surveillance camera placed at the intersection of Piccadilly Circus, London. The 12 hours’ footage was done within two consecutive days with both day and night scenes in a single positive without rotating. This online crowdsourced art is open to the global public who has internet access. It is an open anonymous online environment where users can annotate anything without advance sign-up.

Method

The annotation feature is the unique and fundamental part of Exhausting a Crowd; therefore, JavaScript code was written to extract the annotative data from the website. In the spirit of open source, you are welcomed to use this code attached in the end of this document, if you are interested in doing research on this website. 

For this project, As there are 12 hours’ time frames to record data, a total of two hours of recorded data was selected for analysis in this paper, which is one-hour day scene, 3:00 pm – 4:00 pm, and one-hour night scene, 11:00 pm – 12:00 am. The two-hour data was recorded at a different time with a 5-hours difference because there are different amounts of activities and passersby between the day scene and the night scene.

A total of 8511 data was generated from the website, and there is not much difference between the afternoon scene, 4454 data, and the evening scene, 4057 data. Thematic analysis is a method for identifying and reporting patterns or themes within the dataset, and it allows a wide range of flexibility to sort the data as there is no clear agreement on how to do the analysis (Boyatzis, 1998).

Findings

Surveillance and Privacy

Surveillance is displayed in multi-layers in the Exhausting a Crowd. The bottom layer of the surveillance targets the passersby who are being filmed by the street surveillance camera. This group of participants has no awareness of this experimental surveillance and have no explicit or implicit objection and action to reveal their privacy concerns. Therefore, they are passive participants and experimental subjects in this online crowdsourced art.

The second layer of surveillance is the most important and significant one to analyze in this online crowdsourced art. It involves the digital age’s core characteristics – technological infrastructure (internet access, footages transformation, etc.) and user participation. This layer is also where all the data from. Among those 8511 data, 72 data explicitly express the surveillance and privacy concerns. As online users who participate in annotative are aware that Exhausting a Crowd is an artwork, they often comment creatively, taking into the perspectives of passersby (experimental subjects), which implicitly reveals the message of surveillance and privacy that this artwork tries to deliver. Comments like “stop looking at your phone punk” indicate the surveillance feature of the online users and implies the digital culture nowadays. 

The online users are not only surveilling the passersby, but the artist and other anonymous users are also surveilling them. They can capture and save/share the annotative information. Therefore, the third and the top layer of the surveillance can be divided into two hierarchical structures. The first one, which also is the less powerful one, is those anonymous users who monitor the annotative and can exact information using technological tools or skills to capture, save or share information from this online crowdsourced art. The other level, which is also the highest one, belongs to the artist, Kyle McDonald, who is the creator of this art and can change the direction or alter this art project’s mechanism and participation.

Anonymous Linguistic Expression

If we say that surveillance and privacy are the core messages of this online crowdsourced art; then, the anonymity feature is the fundamental platform that enables users to express their concerns freely. Within the anonymous, crowdsourced online environment, annotative comments tend to be very neutral and creative, referring to experimental objects/subjects descriptions, popular culture, and emotional expressions. From a brief look through the data, describing experimental objects or subjects, in which items refer to the architectures, buses, infrastructures, restaurant, and subjects refer to human beings, makes up most of the data’s expression style. To be more specific, in every 20 data on a page, about 15-16 describe experimental objects or subjects. As the footage contains so many details and activities happening in the intersection in London, some online users also make the association of the places and passersby with popular culture, and annotate on the footage: “MAD MAX” related annotations appear 25 times, “fast and furious 8”, which also appears as an advertisement on the bus; Harry Potter related comments appear six times, “VOLDEMORT,” “I WILL FIND YOU AND KILL YOU HARRY POTTER”. Last but not least, other aspects of the linguistic expressions in Exhausting a Crowd are emotional expressions. Some online users use this anonymous space to express their emotional status.

Discussion and Implication

Privacy is the perceived ownership of one’s personal information and control over how that information is handled (Petronion, 2011; Laufer & Wolfe, 1977). From the lens of this working definition, we can identify that participants in Exhausting a Crowdeither, who are under various surveillance levels, have no control over their personal information. The ‘watchers’ in the higher hierarchic order in this online crowdsourced art know their whereabouts.

The anonymous online environment has often been treated as a place that triggers malicious behaviors or offensive, nasty comments. However, such a rationale seems not applicable in Exhausting a Crowd. Researchers have found that incivility targets content that relates to an individual’s ideological beliefs (Hwang, et. al, 2008); the author’s tone (Price, Nir, & Cappella, 2006), or the online content viewpoints (Phillips & Smith, 2004). And yet, Exhausting a Crowd is an open-end online platform showcasing architects, activities, and passersby without any distinguishable topics that relate to one’s ideological belief, viewpoints nor does it contain a particular tone. Hence, such an open online art exhibition provides a civil environment for participants to engage only with the art themes and express their creativity. The openness feature in both the context and access to this art project helps to widen participation and diversify creativity (Literat & Glăveanu, 2016).

My Other Relevant Work on Online Privacy 

Gambino, A., Kim, J., Sundar, S. S., Ge, J., & Rosson, M. B. (2016, May). User disbelief in privacy paradox: Heuristics that determine disclosure. In Proceedings of the 2016 CHI Conference Extended Abstracts on Human Factors in Computing Systems (pp. 2837-2843): https://dl.acm.org/doi/abs/10.1145/2851581.2892413.

As the research assistant, I participated in the data collection and analysis part - held in-depth focus group discussions and analyzed and coded the data in the constant comparison approach. Our preliminary results yielded common themes on people’s privacy heuristics. 

Findings

Positive Heuristics

Heuristics that increase information disclosure behavior

Gatekeeping Heuristic: When a system has many layers of access, it makes me feel safer
Safety Net Heuristic: I am safe because a third-party (e.g., Visa or PayPal) has me covered
Bubble Heuristic: If I am in a protected network, I am safe
Ephemerality Heuristic: If information is not stored but disappears, then it is safe (to share)

Negative Heuristics 

Heuristics that inhibit information disclosure behavior

Fuzzy-boundary Heuristic: My information may be shared with third-parties, therefore it is unsafe
Intrusiveness Heuristic: If something arrives unsolicited, it is unsafe
Uncertainty Heuristic: If I don’t understand the interface or how it works, it is unsafe
Mobility Heuristic: That which you carry around is inherently unsafe

---

Data Scraping Code

+

var SECONDSINHOUR = 1000 60 60; var numSeconds = 0; var notesAry = []; var startTime = new Date(); var endTime;

// select the target node var target = document.getElementById('notes');

// create an observer instance var observer = new MutationObserver(function(mutations) { mutations.forEach(function(mutation) { var changeData = {};

if (mutation.nextSibling === null) {
        // console.dir(mutation); // do nothing
    } else {
        // console.log(mutation);
        console.log('TIME: ', numSeconds, 'TEXT: ', mutation.nextSibling.innerText);

        changeData.id = mutation.nextSibling.id;
        changeData.message = mutation.nextSibling.innerText;
        changeData.seconds = numSeconds;
        changeData.timestamp = startTime.setSeconds(startTime.getSeconds() + numSeconds);

        notesAry.push(changeData);
    }

}); });

var timer = window.setInterval(function(){ numSeconds++; if (numSeconds === SECONDSINHOUR) { // STOP timer window.clearInterval(timer); observer.disconnect();

console.log('---------- STOPPED RECORDING ----------');
    var textData = JSON.stringify(notesAry, null, 2);

    endTime = new Date();

    console.log('---------- START DATA: ' + startTime + ' ----------');
    console.log(textData);
    console.log('---------- END DATA: ' + endTime + '----------');
}

}, 1000);

// configuration of the observer: var config = { attributes: true, childList: true, characterData: true };

// pass in the target node, as well as the observer options observer.observe(target, config);

---