Data Selfie is a Chrome browser extension that tracks people while they interact – click, type, scroll, dwell – with content on Facebook and that predicts various personality traits based on that data. I used it for two weeks (April 18 - May 1, 2016) and transformed the results into a time-based data visualization that is presented as a projection on translucent material to represent on the one hand the ‘glassy’ consumer and the transparent person, but also on the other hand the lack of information transparency.
Internet users deliberately create different identities on the different platforms they use. They control how they want to be seen. Data Selfie addresses the curiosity about how the data profiles they make compare to the ones made about them by machines at Facebook, Google and Co. – the only profiles they never get to see, but unconsciously create.
This project aims to provide a personal and differentiated perspective on data mining, predictive analytics and our online data identity – including inferred information from our consumption. In our data society algorithms are omnipresent and seemingly omniscient and omnipotent. We can hope that Big Data will solve all problems of humanity or fear a future of total control and surveillance, without freedom. One thing is certain, Big Data algorithms increasingly define our lives. Therefore, it is important – especially for those who are indifferent to this issue – to be aware of the power and influence of our data on ourselves.
Can our data selves help us understand ourselves?! Can it really?
The concept behind Hang’s Data Selfie is one we can all relate to in the growing digital data mining age. I think the angle Hang is taking – the human element that writes the algorithms that create meaning from the uncountable bits of data collected – is a unique and important one. Most people assume that computers are unbiased accumulators, but by showing the different personality profiles built through different algorithms, Hang is highlighting the variability of the profiles built about us. I think in presenting the project, this element should be more prominently displayed, though the prototypes are necessarily limited by having only one projector. The point that the personalities amalgamated and displayed are to be compared to one another, and not, in fact, to the real person herself, would be more obvious that way.
Hang’s computer code aesthetic unmistakably mirrors the back-end algorithmic data-mining concept, and the green lines connecting the actual data – articles, posts liked, etc. – to the weighted analysis pieces clearly delineate the relationship between the two as well as provide an idea of the connections at a glance. As of now, the personality profile generated by the algorithm is output in plain text, and in green. The green differentiates it from the data and initial analysis slightly, but could definitely be emphasized more to more clearly indicate its importance to the viewer. We also discussed the idea of switching the output from plain text entirely and using the actual data percentages formed through the analysis to more closely mirror the rest of the project – both the data and initial analysis projected above the profile and the over-arching concept of computer-formed personality profiles.
Another aspect Hang brings up as a concept behind the project is the use to which companies and other entities will use these profiles – and how those uses are affected by the personality predictions of such varying algorithms, and in turn how those uses affect our daily lives. If she decides to push further in this direction, some part of the project will need to address that next step of the data analysis, with research into the field to make sure it is an accurate portrayal (such as what loan application cutoffs banks use, or what companies analyze for employment opportunities, to go above and beyond the normally thought-of curated advertisements). Or, as she currently has the project, to leave this use of varied and flawed profiles to the imagination of the viewer.
Overall, I have complete confidence in Hang’s concept, use of the technology, process, and production. I think it will end up being shown to great effect with the projections and milky glass panes. I am very interested to see how the different algorithms will shape separate personality predictions based on the same input and analysis of sentiment, entities & subjects, and overarching concepts of Facebook engagements and interactions.
Transparency: German court rules against use of Facebook "like" button on website
A note to transparency and how our online trails are used or rather who actually has access to our data.
“The court found in favor of the North Rhine-Westphalia Consumer Association, which had complained that Peek & Cloppenburg's Fashion ID website had grabbed user data and sent it to Facebook before shoppers had decided whether to click on the "like" button or not.
"A mere link to a data protection statement at the foot of the website does not constitute an indication that data are being or are about to be processed," the court said.
Germany, Europe's biggest economy, is one of the world's strictest enforcers of data protection laws and its citizens have a high sensibility to privacy issues.”
Source: Reuters, further reading Der Spiegel (German)
I also just had an interview (together with Regina for the NYC Media Lab Combine program) with Teaching Privacy’s* Gerald Friedland and Julia Bernd which is supported by the International Computer Science Institute. Gerald, being German, also confirms that these privacy concerns seem to be influenced by culture.
I have created a Chrome browser extension that tracks myself while I interact – click, comment, hover, scroll – with content on Facebook and that predicts various personalities based on that data. I have used it for 2 weeks (April 10 - 24, 2016) and visualized this as a time-based data visualization that is presented as a projection on glass to represent the 'glassy consumer‘ and the metaphor of transparency.
Internet users purposefully create different identities on different platforms they use. They curate, manage, essentially control how they want to be seen. Data Selfie addresses the curiosity of: What about the data profile that is made of us by machines at Facebook, Google and co., the one profile we never get to see, but unconsciously create?!
Data Selfie aims to provide a personal and differentiated perspective on data mining, artificial intelligence and your online data identity. So many details and assumptions can be made from our consumption and depending on the subjective rules the algorithm architect designs, there will be different outcomes. Specifically, the seemingly endless stream of mined information will include what content I look at and read, in what languages and in what tones, what I type, how much time I spent on Facebook, what I like and comment on, who I interact with, who I stalk. This data will be visually and procedurally connected to the different, inferred identity profiles. No imagery is used in this visualization to emphasize how these processes are overseen by machines rather than human beings.
In the information age data dictates the economy. We can either hope that big data will solve all problems of humanity or fear a future of total control and surveillance without freedom. Data Selfie is born out of the belief that both have to be taken into consideration. One thing is certain, data is increasingly defining our lives. Therefore, it is important for especially those who are not hopeful nor fearful, who are indifferent to this topic to be aware of the power and influence of data of ourselves on ourselves. Can our data profiles help us understand ourselves?! Can it really?
Transcript of my presentation on March 3, 2016 and production calendar
- previous projects and their connection to thesis
- The Nature of News (MS1) and ham$ta (MS2) are both results of my thoughts about media, specifically social media and the influence on our lives in terms of how we see the world and how we learn new things
- since a lot of people’s lives more and more take place on the internet (Edward Snowden says living in Russia is not that difficult or a burden because he lives online), I wonder what can all this data tell us that we don’t know yet.
- can it help us understand ourselves better?
- does Facebook know me better than I know myself?
- to begin answer these questions, I am looking into the domains of self tracking, data mining, predictive analytics, data/online identity
- this is interesting and important, because Facebook doesn’t share any of the data they have about yourself
- so, I am creating a chrome extension to gather that data, that Facebook tracks and uses to target ads to me
- essentially I am hacking their front-end infrastructure to retrieve data that is actually mine.
- and I will use this tool for 2 weeks and create an art installation in the form of a projection on glass as a result
- I explained what elements, what analysis will be included (please see presentation)
- my main precedents for this digital data display (basically data visualization) are Code Names of the Surveillance State by Trevor Paglen (2001 to present), Internet Cache Self Portrait by Evan Roth (2014), Listening Post by Ben Rubin (2002); all interesting records of data, somehow secret or aggregated
- the glass is a core element, while the rest of the current installation mock-up might be changed
- transparency is a core concept in this project in different aspects, one we are “glassy consumers”, transparent to companies, two companies are not transparent with us
- not only is our data not shared with us, we don’t have any way of knowing what assumption are made about our personality not even to mention how they have arrived at these data profiles
- it’s not hard to imagine how these practices. when used more and more in even more aspects of our lives, can have negative implications, not for all, but for some, issues about privacy and discrimination
Production Calendar
(revised March 15)
Week 6 _ March 08 - March 13
> Optimize data gathering + explore visualization/display
Internet users purposefully create different identities on different platforms they use. They curate, manage, essentially control how they want to be seen. Data Selfie addresses the curiosity of: What about the data profile that is made of us by machines at Facebook, Google and co., the one profile we never get to see, but unconsciously create?! Can our data profiles help us understand ourselves?! Can it really?
Data Selfie aims to provide a personal and differentiated perspective on data mining, artificial intelligence and your online data identity. This project consists of two parts: The first is a Chrome browser extension that tracks the user while she or he interacts with content on Facebook and predicts your personality based on that data; the second is an installation that displays my data, the result of using the tool for two weeks. The transparent material on which the data visualization is projected represents the 'glassy consumer‘.
So many details and assumptions can be made from our consumption and depending on the subjective rules the algorithm architect designs, there will be different outcomes. Specifically, the seemingly endless stream of mined information will include what content I look at and read, in what languages and in what tones, what I type, how much time I spent on Facebook, what I like and comment on, who I interact with, who I stalk. This data will be visually and procedurally connected to the different, inferred identity profiles. No imagery is used in this visualization to emphasize how these processes are overseen by machines rather than human beings.
In the information age data dictates the economy. We can either hope that big data will solve all problems of humanity or fear a future of total control and surveillance without freedom. Data Selfie is born out of the belief that both have to be taken into consideration. One thing is certain, data is increasingly defining our lives. Therefore, it is important for especially those who are not hopeful nor fearful, who are indifferent to this topic to be aware of the power and influence of data of ourselves on ourselves.
A New Tool for Tracking the Use of Personal Data on the Web
“The web today is a big black box, what’s needed is transparency.”
“That leaked targeting information can potentially be used for all sorts of purposes. It can be used for discrimination. And it is a very hidden kind of discrimination.”
- Roxana Geambasu, assistant professor of computer scientist at Columbia University
Lee Tien, a senior staff lawyer for the Electronic Frontier Foundation, said XRay seemed promising and a way to help shift the terms of trade somewhat between consumers and data harvesting companies. “Seeing what they see,” Mr. Tien said, “is a first step toward some kind of a balance.”
Further reading about Sunlight, the successor to XRay, by the same researchers at Columbia here.
What if we would pay for Facebook and co. to not track us? Would that create a different set of problems? Will it then become a question who can afford to protect their privacy?
ellievsbear
titsay
pixelskylines
izzy806
blakekathryn
gummygunk
discoholicmusic
germansingergirl
annaxmalina