Online Correspondance / EU Digital Futures
A collaboration with postdoctoral researcher Martin Skrodzik and artist Tiz Creel.
DATA IS EVERYWHERE
The experiment intent is to infer personal biometric data based on seemingly random questions. Two of the ten questions were formulated and designed to enable us to mine information without the participants knowing.
We chose to guess the participant’s age.
TEST IT YOURSELF -------->
79
Responded
The participants respoded to more than 5 questions. 1 participant responded no questions.
questons
most
People
80
We recived 82 resposnes witch 2 of them where duplicates
Responded
15
Responded
questons
all
PART 1: RANDOM QUESTIONS
The Chinese zodiac sign reduces the possible years the participants were born into seven options.
1. What is your Chinese zodiac sing? (69 responses)
Of all your internet lifetime, how much could potentially reveal personal or sensitive information?
How much could data reveal your location, recurrent places, and employment details?
What about the information that is not implicitly there?
Different generations have used various social media throughout their lifetime
We could attribute many of them to a particular generation. Additionally, there are also age groups when it comes to current social media usage.
2. What social media did/do you use? (79 responses)
Who owns your data?
The Chinese zodiac sign reduces the possible years the participants were born into seven options.
3. What is your favourite Album? (70 responses)
The Chinese zodiac sign reduces the possible years the participants were born into seven options.
4. In what year did you have your first kiss? (70 responses)
All your uploads, tags, conversations, all the statements that perhaps are not true to you anymore, or all the times you acted out of character on social media.
Data is shared privately, and publicly.
Data that others shared.
THE MODEL
We will infer the age of the participants with only two factors: the year of the first kiss and the Chinese zodiac sign.
According to multiple studies, most people have their first kiss between 12 and 15 years old.
- The Chinese zodiac signs are repeated every 12 years, giving us leverage to expand the average of the first kiss to between 10 and 21 years old.
1. We created a model that calculates the first kiss age in all seven possible years according to the given Chinese zodiac sign.
2. The model selects a year in which the age of the first kiss is between 10 - and 21 years old.
The formula predicts the age of the participants with 90% accuracy.
ERRORS
The 10% failure is in all cases of people who had their first kiss outside the selected gap between 10 - 21 years old. Therefore the model is unable to compute the age.
Another potential discrepancy: the Chinese zodiac sign runs from February 1, 2003 – to January 21, 2004, which can create problems with the western understanding of the new year. However, it seems to not affect our sample.
Every prediction with one year difference is computed as a valid response. There is the leverage of one year relative to when the test was responded (in this sample is the 30th of June) and the date and month the participant was born. An interesting outcome of this discrepancy is that we can know if the participant has not yet celebrated their birthday this year.
“The questions. What was your first favourite album?” And, “what social media do/did you use?” Are meant to be used as confirmation bias and inspect the responses that the model could not compute.
Disclaimer 80 participants 53 shared the essential data with participating in the age prediction.
The model was done with the help of excel wizard Maria Skibinska
PART 2: RANDOM DATA QUESITONS
As part of our investigation, we wanted to see weather the participants stand regarding their data.
5. Do you have something to hide in your data?
(79 responses)
6. Would you agree to have all your data publicly accessible?
(79 responses)
7. Did you change your answer to question 5 after answering question 6?(79 responses)
8. Write a text of up to 500 characters on the statement "Data is everywhere" (50 responses)
PART 3: "DEMOGRAPHICS"
9. If you want to take a chance in winning a 50$* Amazon gift card for answering this survey, enter your email address here.
(20 responses)
Disclaimer : * As for the answer to question 9, please note that we are not actually giving away a 50$ amazon gift card. We just wanted to see whether people would enter their email addresses. The chances are exactly 0% for winning a gift card.
We wanted to see how many people would give away their email after the rather suspicious questions on data and the clear disclaimer at the end of the survey that states that there is no gift card and the question is just a test.
Of 83 participants:
-
12 People who did give away their email, which naturally implies, that they did not read the disclaimer.
-
6 People gave fake emails, we could assume that they read the disclaimer. As we have one email that reads: hahaIreadthedisclaimer@gmail.com
-
Others: ha@haha.verygood, dollars@cents.com, anythingelsetowin?@aol.com, ocp1991@thatsmyyearofbirth.com
-
-
65 People who did not give away their email, can just hope that they did read the disclaimer.
10. For demographics, would you please tell us your age? (74 responses)
We asked the age for “demographic purposes”, but actually, for us to confirm our predictions and test our model accuracy.
Conclusion
We want you to think, how easy it would be to extract sensitive personal information from your public profiles (including posts, media, descriptions).
How many online tests, quizzes, questionaries have you filled out throughout the years?
How many terms and conditions have you read?
How many conditions have you ageed to?
How many companies have your data?
As internet users, how much do you really really really care about your data?
The goal of this experiment was to see how easy it is to infer data from seemingly innocent questions, information that we regularly share on social media.
Make your conclusions.
