Return to Project Overview

Interactivity Project: Experiencing an Alien Virtual Environment

                How do we experience an alien virtual environment? A virtual environment is simulated reality. A form of virtual environment is constructed through human interaction with computers and the internet. The environment we create is made familiar through personalization. What one experiences is not the same as another. It is “alien.” When we access the internet through someone else’s environment, we alter our experience. The question is “how?” The “Alien Environment Project” was designed to allow users to experience contrasting alien internet environments so that they might become more aware of how personalization impacts their information experience. The project began with an analysis of internet data collection followed by the construction of two contrasting archetypical characters who then became the source of their own virtual environments. The resulting products were compared and analyzed. Users were invited to enter the environments and experience alien perspectives.

Highly personalized virtual environments are created for individual internet users. When we access the internet, we are both recipients of data and distributors of it. We engage with information and we leave behind trace evidence of our engagement. According to Eli Parser’s 2012 book, “The Filter Bubble” the data that we generate is collected and analyzed by complex algorithms. These algorithms then take an active role in shaping the information that we receive.[1] The data that is filtered through each “Filter Bubble” comes from a variety of sources. Many of the data sources are known and understood. Some are obscure. Social networking and commerce websites keep track of personal data that users voluntarily enter to create their personalized experiences. This data is known to and theoretically controlled by the user (the question of internet privacy further complicates the issue).  Data is also collected from other sources. Pariser claims that Google developers admit to a system of data collection which contains “57 signals” which are used to analyze and predict user tendencies. These “57 signals” have not been published. They are not known and cannot be directly controlled by the user. This means that individual environments, while personalized, are not necessarily a product of individual desires. Instead, they represent an algorithmic interpretation of desire.

Character profile requirements for the project were determined based on both direct and obscured internet data collection.  Many websites contain user profiles containing basic demographic information voluntarily entered by the user. Sometimes this is basic information like “name, age,” and “location.”  Social networking sites collect more detailed data that can include everything from “relationship status” to “annual salary.” The information requests from multiple websites were compiled. The websites accessed included representatives from the most common sites related to the following internet uses: direct communication, networking, information research, entertainment and commerce. This includes but is not limited to: Google, yahoo, face book,, YouTube, Amazon, Wikipedia and Netflix. The list of questions was thinned down by excluding information specific to only one of the sites as well as information that would require significant conjecture to provide. The resulting list was then used as a foundation to create character profiles.

The determination of initial character traits were carefully made. The project was limited to two contrasting characters (additional characters could be created in the future). The characters were not based on real people. Nor were they designed to promote stereotypes designed to homogenize groups or demean individuals. The characters are archetypal characters created to represent differing segments of society and therefore create contrasting environments. There was significant debate about the nature of the first characters. Initially, the characters were to represent an extreme majority and an extreme minority. However, response to the US Presidential election suggested a different path. The political divide has become a political chasm. The contrasting points of view are extreme and yet they represent roughly equal portions of our society. This became the starting point for character development. The characters were selected to represent opposing political views of the majority populations. This provided a needed central focus to guide decision making when statistical analysis and data collection were not wholly sufficient sources.

The characters profiles were completed using careful research. These profiled were utilized to provide direct information to generate the internet environment. The process is described below. The information is deeply interconnected. As such, it was not gathered in linear order. It was assembled like puzzle pieces based on total requirements.


The surnames names were chosen based on location and age. The most common surnames of each state that were different from each other. Maine lists, Smith, Brown and Johnson. Alabama lists Smith, Williams and Johnson.[2] The first names were created by according to the US Census database[3] record of the most popular gender-specific first names for the year of birth.


The age of the characters was determined by gender and voting patterns. Income, 0ccupation and education were secondary factors.

Determination of age began by accessing state profiles from US census bureau. [4] However, both Alabama and Maine have similar age and gender demographics. Contrasting characters suggested contrasting generations. One character needed to be significantly older than the other. The given assumption that one character voted for Clinton and one for Trump provided a solution. Since young women were more likely to vote for Hilary Clinton,[5] it was determined that the younger character would be the female.

The characters employment and end education histories were also considered. The male character’s occupation and yearly income suggest an age close to retirement. The female character’s occupation required minimum education levels which determined a minimum age of 22, adjusted up to allow greater income and experiences.

With no specific justification (for kicks), the characters share the same birthday, November 20.

Race/Sex/Gender/Sexual Preference/Marital and Family Status

The race of both characters was determined according to the majority population in each state. Maine is a whopping 95.9% white. Alabama is only 69.8% white, with 26.8% black.[6] Since only 2% of Republicans are white[7] and Alabama is a Red state, it was determined that the character from Alabama would also be white.

It was determined that the characters would be of opposite sex to create contrast.

Gender and Sexual Preference were statistically normative.

The family statistics for the both represent the average statistics according to US census data. The female statistics also represents the trend towards increasing age of first birth.[8]


The region was selected based on voting patterns and aesthetic balance. The initial given required a red and blue state. The decision to use Alabama was simple. It is my home state. The decision to use Maine as the blue state allowed me to use my home state and my new home.

There are some problems with selecting Maine as the second state. It is a politically divided state with a large conservative streak. However, for the purposes of this project it was decided that the character created would still remain an accurate representation of the democratic population of Maine.  Cumberland  County was specifically chosen because of the election results. It had the most democratic voters.[9]

The housing was based on I selected housing based on race, income and employment. The neighborhoods were searched for demographic and income data. Addresses were provided by vacant houses on


Occupation was predetermined by the 4th largest employer in each State and County of residence. The salary was looked up according to occupation, age and likely experience level.

The results were Hyundai and Inspiria Health Services[10]. 20,000 was deducted from the nurse salary because a 25-year-old should be entry level.

Religious Affiliation

Research was not conducted on religious affiliation in Alabama. This was based on personal knowledge.

Maine was ranked the 48th least religious state in two different polls.[11]

Education Level

Education level was determined by state statistics as well as occupation.

A nurse requires a college degree. The University of Maine is the most attended university in the State of Maine.

The Hyundai plant does not require a college education to work in quality control. State statistics show high-school education as the most common education-level. Based on age, location and race, Robert E. Lee High School would have been a likely choice for a white, middle class boy in 1971. (No, the name has not been changed.)

It should be noted, that race is a determining factor for education in Alabama. Despite the well known integration laws of the Civil Rights Movement, many schools remain racially segregated. The complexities of the County v, City School systems in Alabama in general and Montgomery in particular are both the result of and the perpetuators of racial segregation. This has significantly improved over time. Current school systems are better and more integrated. I attended high school at a school which was almost exactly 50/50racial mixing. This has become much more common.

Political Interests

Each character was given political interests. The political issues came from 5 most interesting measures on Maine and Alabama ballots. “Interesting” was determined roughly based on the number of time the results were mentioned in search results: ” political issues in Maine/Alabama[12]” The listed voting record is based on the majority results in the given counties. [13]

Hobbies and Interests

Hobbies were necessary to provide potential search terms and to select Apps and shopping patterns. I selected generic interests based on my own general knowledge, local club listings, newspaper ads, University clubs and newspaper descriptions.  Travel descriptions of each state and Craig’s List personals were also consulted.

Family Issues

Family issues were selected to add depth to face book profiles and fodder for search results. The characters issues related to their age and circumstance. The age of the male suggested the likelihood that at least one child would be in college. The female was given an ailing parent. The ailment is based on the most commonly diagnosed condition in the US.[14]

These are the resulting character profiles. They were used to create all applicable project profiles.

Character Profiles

James Williams

November 20, 1953







South East



Montgomery County

3329 Sommerville D. Montgomery, AL 36111

Hyundai Manufacturing

Quality Control Officer




Church of Christ

Robert E Lee High Class of 71

ü Voter Registration IDs

ü Prohibiting Reallocation of Parks

ü Right to Work Amendment

ü Etowah Law Officer Supervision

û Toll Roads


Football (Roll Tide)


1 Child in College, UoA


Date of Birth





Marital Status






Street Address

City, State & Zip code



Yearly Salary

Home Ownership

Political Affiliation

Religious Affiliation


Political Issues





Hobbies and Interests



Family Circumstances

Ashley Brown

November 20, 1991







North East



Cumberland County

397 Island Ave.

Portland, ME 04108

Inspiria Health Network



Lives at Home


Not Affiliated

University of Maine Class of 13

ü Medical Marijuana

üGun Control

ü Increased Minimum Wage

ü Income Tax for Education

ü Ranked Choice Voting




Father with Hypertension

Manipulating obscured data is more difficult. Google has not publicly confirmed the “57 Signals” it uses to analyze user data. Nor have similar companies publicly confirmed their data. Computer theorists have speculated. Rene Pickhardt is a PhD student at the University of  Koblenz. He made a list of the signals he speculates that Google may use.[15] His list has been cited numerous times throughout the internet but not confirmed by Google or peer review. However, it provides an interesting list of actions that can be recorded. It is also possible to consider the types of data stores by internet cookies.[16] Cookies store user data that so that websites can personalize experiences.  Google maintains a database of common search phrases. Common search terms can be accessed on terms on Google trends[17]. The character profiles were used to create a likely search history.

The virtual environments were assembled using iPads. The iPads were reset to factory settings. Location services were turned off in an attempt to remove local influences. Each iPad was set up as a new iPad belonging to the profiled characters. The operating systems were updated. Apps were installed, including Face book, Google Suite, Amazon and Kindle as well as character specific Apps related to hobbies and interests. Character profile data was entered in the available Apps. Finally, a

Character Login and Password Information

James Williams



Best Friend in High School


Model of First Car

75 Ford F-100

Street where you grew up

Wares Ferry Road





Apple ID








Face book



Ashley Brown



Best Friend in High School


Favorite Children’s Book

Harry Potter

First Beach




search history was created. After each search, the first three non-advertisement links were clicked on each page to produce a web of related searches. This process  was repeated at least 3 times a day over the course of at least three days.

Character Search Terms

James Williams

Housing Costs in Montgomery Alabama

Hyundai Retirement

AmSouth Bank Online Banking

Dalriada Church of Christ

University of Alabama

Republican Candidates

Donald Trump

Election results

Did trump win

Obamacare repeal



Voter Registration IDs

Prohibiting Reallocation of Parks

Right to Work Amendment

Etowah Law Officer Supervision

Toll Roads Stink

How much money to retire


Fish Stories

Fish comics

How to catch bass

Best bass boats

Alabama Football schedule’

Roll Tide

Radial Arm Saw

Dewalt or Makita?

Best Scroll Saw

Ben Carson


Credit Karma

Alabama Attorney General

Ashley Brown

Apartments in Portland Maine

Roommates in Portland Maine

Inspiria Health Network

Camden National Bank Online Banking

University of Maine

Democratic Candidates

Hillary Clinton

Trump impeachment

Election recount

Electoral college


Winter temperatures 2016

Medical Marijuana

Gun Control

Increased Minimum Wage

Income Tax for Education

Ranked Choice Voting

What to read next

Best books of the year

Cross country skis

Best cross country in Maine

Magnolia Story

Thank you for Being Late

Bernie Standers: Our Revolution

The Boys in the Boats

Hypertension Life Expectancy

Hypertension Diet

Heart Disease Genetics

Amy Shumer




The construction of the environments had complicating factors. These have been listed for consideration:

  • The iPad stored data should have been completely deleted when reset. However, there were occasions when old IMRC login and password were requested. A former students video and camera files also reappeared during a Google Sync despite being logged in under the character id. This issue was never reconciled and suggests that the obscured data remains present.
  • Location tracking could not be fully disabled. While “Location Services” was disabled on The iPads, the location was still available once wifi was accessed. A VPN was considered but it was determined that use of the VPN would be more destructive than helpful to the construction of personalization.
  • The absence of usable credit card data limited the detail available in E-commerce transactions.
  • The project was conducted over a 3-day period. Many internet environments are created over the course of many years of use. This factor, while not necessarily limiting, should be considered when evaluating the outcomes.







[1] Pariser, Eli. The Filter Bubble: How the New Personalized Web Is Changing What We Read and How We Think. Penguin Books, 2012.

[2] “What’s the Most Popular Surname in Your State? – Ancestry Blog.” Ancestry Blog. 2015. Accessed November 20, 2016.

[3] “Social Security.” Popular Names by State. Accessed November 20, 2016.

[4] “Selected Characteristics of the Citizen, 18 and Older Population.” US Census Bureau. 2016. Accessed November 20, 2016.

[5] I lost my source. Find it again asap.

[6] “Selected Characteristics of the Citizen, 18 and Older Population.” US Census Bureau. 2016. Accessed November 20, 2016.

[7] Gallup, Inc. “Democrats Racially Diverse; Republicans Mostly White.” 2013. Accessed November 20, 2016.

[8] “Average Age Of First-Time Moms Keeps Climbing In The U.S.” NPR. Accessed December 11, 2016.

[9] “Maine Election Results 2016: President Live Map by County, Real-Time Voting Updates.” Election Hub. Accessed November 21, 2016.






[15] “What Are the 57 Signals Google Uses to Filter Search Results?” Ren Pickhardt RSS. Accessed November 20, 2016.

[16] “Cookies And Privacy FAQ.” Cookie Central. Accessed December 11, 2016.