1. Introductions
1.1. As information itself becomes the largest business in the world, data banks know more about individual people than the people do themselves. The more data banks record about each one of us, the less we exists - Marshall Mcluhan
1.2. The social Data Revolution
1.2.1. How can we ensure that data are for the people?
1.2.2. Give to get - people create and share social data.
1.2.3. Social data doesnt end with you. YOu create and share data about the strength of your relationshps with family, friends and colleagues through your comminication patterns.
1.2.3.1. you leave digital breadcrumbs everywhere
1.2.3.2. even in digital thermostats!
1.2.4. Principles of Post Privacy Age
1.2.4.1. efforts to safeguard citizen's interest
1.2.4.1.1. fair use of information
1.2.4.2. Principles of transparency and Agency provide a means of protecting us from the misuse of social data while increasing the value that we are able to reap from them.
1.2.4.2.1. Transparency encompasses the right of individuals to know about their data: what is is, where it goes, and how it contributes to the result the user gets.
1.2.4.2.2. Agency encompasses the right of the individual to act upon their data.
1.2.4.3. Balancing the power
1.2.4.3.1. Information is at the center of power
2. Chp1:Becoming Data Literate
2.1. 6 rights
2.1.1. Right to access YOUR privacy data
2.1.1.1. Bild zeitung newspaper
2.1.1.2. building tools to allow people to access their data is the next most important!
2.1.1.2.1. make the implicit explicit
2.1.1.2.2. make tools to make meaning out of the data
2.1.1.3. who has the most power over you as a CONSUMER?
2.1.2. right to be remembered
2.1.2.1. Right to be forgotten - delinking
2.1.3. Right to inspect the data refineries
2.1.3.1. need for benchmarks
2.1.3.1.1. How secure is the data?
2.1.3.2. look at 2 factor authentication
2.1.3.3. stickers for the tradeoffs
2.1.4. Right to Amend
2.1.4.1. empower people
2.1.4.1.1. let users annotate anything
2.1.5. Right to Blur
2.1.5.1. Geolocations/ best example
2.1.5.2. consumer should be in charge of the granularity
2.1.5.3. People need to understand the tradeoffs.
2.1.5.3.1. pizza can't be delivered to your doorstep if you 'blur' your gps coordinates
2.1.5.4. Doxing is the public, digital release of a person's private information without their consent, usually to exert some kind of power over the dox-ee. But the practice hasn't always been digital
2.1.5.5. Werner "differential privacy ideas"
2.1.5.5.1. Data for the people
2.1.5.5.2. Did you smoke pot? if you did not, you can say No, or Yes.
2.1.5.5.3. so the police can not come back and arrest you
2.1.6. Right to Port
2.1.6.1. make it easy for a user to transfer the data (port the data)
2.2. What is your worth to them?
2.2.1. we increasingly rely on data refineries for more important decisions.
2.2.2. Algorithms find patterns that humans can't see.
2.2.2.1. true meaning emerges when we compare OUR data with OTHER's data.
2.2.2.2. The value depends on how useful the insights are. How the outputs benefits us.
2.2.3. we should ask for control over the output of the data.
2.2.4. exploration vs exploitation
2.2.4.1. the trade offs
2.2.4.2. multi armed bandit example
2.2.4.2.1. The crucial tradeoff the gambler faces at each trial is between "exploitation" of the machine that has the highest expected payoff and "exploration" to get more information about the expected payoffs of the other machines. The trade-off between exploration and exploitation is also faced in machine learning
2.2.4.3. OIl and Data
2.2.4.3.1. apply resources to dig deeper? or shift resources to finding new areas with oil?
2.2.4.3.2. for data, the time and effort of users
2.2.4.4. Fuzzy Suitor problem
2.2.4.4.1. magine an administrator who wants to hire the best secretary out of {\displaystyle n} n rankable applicants for a position. The applicants are interviewed one by one in random order. A decision about each particular applicant is to be made immediately after the interview. Once rejected, an applicant cannot be recalled. During the interview, the administrator can rank the applicant among all applicants interviewed so far, but is unaware of the quality of yet unseen applicants. The question is about the optimal strategy (stopping rule) to maximize the probability of selecting the best applicant. If the decision can be deferred to the end, this can be solved by the simple maximum selection algorithm of tracking the running maximum (and who achieved it), and selecting the overall maximum at the end. The difficulty is that the decision must be made immediately.
2.2.5. Transparency demands
2.2.5.1. users get to see the refinery settings
2.2.6. agency demands
2.2.6.1. users have ability to affect them
2.2.7. Social data may be harder to verify (signal or noise) ? depends on context
2.3. becoming data literate
2.3.1. Literate: able to read materials considered essential for survival.
2.3.2. shed the old mindset of just using data
2.3.3. we must be co-creators of data
2.3.3.1. only way to ensure that data is for the people
2.3.3.2. data is the new oil.
2.3.3.2.1. crude oil needs to be refined, so too does raw data. The value is created by refineries that create new apparatus of social data revolution
2.3.3.3. unlike oil, data supply is increasing, and the cost of refining them is going down
2.3.3.3.1. data can be accessed by many people
2.3.3.3.2. laws and social norm are based on limited supply. But doesnt really apply to data.
2.4. Data Refining Process
2.4.1. Data broker Acxiom in 1969
2.4.2. Amazon - Save Everything Store"
2.4.2.1. 500 attributes for each user- spend a dollar on marketing? or a dollar on price reduction?
2.4.2.2. customer's purchase history often was less predictive of a purchase than the product's relationship to other products (market basket analysis)
2.4.2.2.1. customer's browsing of an item, the queries, clicks and purchases of previous customers were combined an analyzed to suggest substitutes.
2.5. What's your Data worth?
2.5.1. defined by how useful its outputs are for our decision making
2.5.2. Exploration vs exploitation
2.5.2.1. how to strike the balance between exploring the data, and starting to use it. analogy: Oil. if the reserves are drying up, do you spend money to drill deeper (exploration), or use up the available oil, then explore somewhere else? (exploitation)
2.6. Learning from Errors
2.6.1. learn to separate signal from noise.
2.6.2. use user feedbacks to do error corrections
2.7. Turning data into decisions!
2.7.1. descriptive, predictive and prescriptive
2.7.2. american airlines realized that travel agents are most likely to choose the first item in the listing. So they put their most expensive listings on top. Turns out badly later on. When competing airlines found out. The bias stopped when govt prohibited it
2.7.3. unlikely if the data refinery is the users themselves. Agoda example. Ranking was not based on preference of travelers instead of agoda's profit
2.7.4. NASA moon landing. Prediction. ID actions to take next in reaction to the data.
2.8. Experiment! Experiment! Experiments
2.8.1. Data refineries depends on AB testing experiments. Change one variable and compare responses against the control group
2.8.1.1. as social data become ever more integrated into our problem solving and decision making, data refineries will develop products and services in context with great significance in our lives, including health care,and education.
2.8.1.2. time frame ? should it be a day or longer? What if it is for educational purposes??
2.8.2. Predictive analytics help amazon save voer a 1m USD a year
2.8.2.1. Amazon France ERROR (forgot to add shipping) resulted in a tsunami of orders! Hence, turning this error into a feature.
2.8.3. Visits by robots may account for 15% of total web traffic, so you need to filter these out.
2.8.4. Cookie assignment might change over time thus causing the same person to be classified into different groups
2.8.5. Users on iphone vs android users
2.8.5.1. iphone default refresh rate may be set higher. Causing the system to erroneously conclude that iphone users are more heavy users.
2.8.6. Time frame from idea to execution may take weeks before, but now it is measured in hours
2.9. 3 Categories of social data
2.9.1. our clicks
2.9.2. our connections
2.9.3. our context
3. Chp2: Character and Characteristics
3.1. principle of indirect observation. Traces we leave define us. Allows data refineries to predict
3.1.1. from conspicous consumption to conspicuous communications
3.2. brief history of privacy
3.2.1. chimneys as first privacy device
3.2.2. US first amendment 1792
3.2.3. first govt secret ballot in victoria 1856-1896
3.2.4. The right to privacy 1890
3.2.5. Google, FAcebook, privacy is an illusion
3.3. From Walls to Windows
3.3.1. Google search algorithm
3.3.2. Google gmail
3.3.3. Facemash (pre facebook) scraped photos from year books
3.4. david stillwell. big five' traits: openness, conscientiousness, extraversion, agreeableness, and neuroticism—the OCEAN model.
3.4.1. people tend to present an accurate portrait of themselves on facebook
3.4.2. You are what you like APP
3.4.2.1. 88% of the cases, they were able to guess homosexuals: likes of Mac Comestics , Wicked the Musical. For Male, confused after waking up from naps. Wutang Clan.
3.4.3. Photos of YOU.
3.4.3.1. Yan Le Cun in Facebook (Deepface) can tag you.
3.4.3.2. Other software can analyze the bckground of the photo.
3.4.3.2.1. social butterfly?
3.4.3.2.2. lonesome adventurer?
3.4.4. People share data to help data refineries to make insights that help the users.
3.4.4.1. you need to provide accurate inputs to get relevant outputs
3.4.4.2. tradeoff, More Utility vs Privacy
3.4.5. Biocache forces users to verify their identity.
3.4.5.1. authentication based on the way you use your mobile phone.
3.5. On the internet, everyone knows you are a dog.
3.5.1. Sweeney estimated that 87% of american public could be identified given sex, birthday and zip code..Later research showed it to be closer to 63%
3.5.2. AOL anonymized search logs of 658000 users were identified based on their search history (they like to search for themselves and their relatives.
3.5.3. Netflix 100m movie ratings from 480K customers were de-anonymized by Univ of Texas at Austin, Arvind Narayanan and Vitaly Shmatikov. by comparing it with IMDB.com
3.6. What's in a Pseudonym?
3.6.1. BioCatch designs its data collection by forcing users to perform actions that verify their identity without their realizing that this is what they're doing.
3.6.2. Game site's models are able to identify the age of a child within an accuracy of 3-6 months based on mouse movements, because the development of fine motor skills is highly correlated with age in children and young teenagers.
3.6.3. At Amazon, customer reviews were more valuable when they were non-anoymous reviews. The most important factor was the indication that the reviewer purchase rather than just the name.
3.7. Honest Signals
3.7.1. Dating Apps using simple count of clicks and contact messages will reveal the true preferences
3.7.2. Netflix, ratings given were far poor signal than the time you actually watched the video.
3.7.3. Jack'd gay dating site, provides data on a users' reply rate to incoming messages and descriptive statistics (age, ethnicity, body type, etc) about the people he's actually shown interest in.
3.7.3.1. more private than these explicit signals to other individuals are the trail of clicks each person leaves behind while exploring user profiles.
3.7.4. Fridae (sg dating site) what we want changes based on time of day or day of week)
3.8. Calling for accountability
3.8.1. PHone number HIYA whitepages forces transparency using phone numbers
3.8.2. Right to be forgotten
4. Chp3: Connections and Conversations
4.1. The fake Linkedin Rebecca fooled recruiters into approaching her
4.1.1. Darwin's use of duration of a man's friendship as an indicator of the man.
4.2. Your Neighborhood in the Social Graph
4.2.1. Dunbar's 150 friend limit
4.2.2. Mark Z. social graph to refer to how people are connected in Facebook
4.2.3. Pyschiatrist JL Moreno's "Epidemic of Runaways" where the runaways social graph influences other runaways.
4.2.4. MCI's Friends and Family program that attempted to lure away AT&T clients by asking them to list up to 20 numbers that they could call long distance at a discounted rate (20% off)
4.2.4.1. in 2 years, 10 million customers were enrolled in the plan
4.2.4.2. MCI had given their customers a financial incentive to persuade their frequent contacts to switch to MCI
4.2.5. social network may not be balanced
4.2.5.1. it is indicated by the number and direction of interactions
4.2.5.2. Mark Granovetter's the Strength of Weak ties
4.3. The New Social Capital
4.3.1. Feeling of friendship involves sequential revelation of information
4.3.2. AT&T used A/B test to compare success of campaign based on social graph data
4.3.2.1. They marketed to people that had interactions with another person who already signed up for the promo. Homophily.
4.4. Designed for the People with the data
4.4.1. Linkedin was interested in helping people find the weak ties that would assist them in getting to where they wanted to be.
4.4.1.1. Ellen Levy, ex VP at LinkedIn said the real breakthrough came when the company figured out how to design the site for the people who have the data, not for those who need it!
4.4.1.2. Challenge was to deliver services that would inspire its users to create and share more data about their professional social network.
4.4.1.2.1. User's needed more excuses to communicate more regularly.
4.4.1.2.2. Linkedin began sending alerts when users added skills, anniverasies etc.
4.4.1.2.3. It provided forums where users can post articles to showcase their expertise and comment on relevant news.
4.4.1.2.4. Added the option to endorse skills
4.4.1.3. WeChat vs Facebook
4.4.1.3.1. Facebook allows you to see a person's friends, WeChat does not
4.5. Ministry of Social Engineering
4.5.1. Facebook Algorithm requires you to give data to get data. 62.5% in a study of FB News Feed didnt even realize an algorithm was refining what posts they are shown.
4.5.2. Emotional contagion study by FB and Cornell found that emotional contagion does work by manipulating the news algorithm to show more or less posts with positive or negative words expressing emotion .
4.5.2.1. backlash as there was no prior Informed consent
4.5.3. Jonathan Zittrain, law professor at Harvard Civic Engineering Project
4.5.3.1. Influencing People to get out to vote
4.5.3.1.1. ONe group was shown ad reminding them to go out and vote.
4.5.3.1.2. one group was shown the ad which included the names and profile photos of friends who had already voted.
4.5.3.1.3. the other group was only shown informational message
4.5.3.1.4. Effects of these two were measured in 3 ways
4.5.3.1.5. An extra 340,000 people went to the polls that Election Day as a result of them having seen the Social version of the ADs
4.6. The Value of Trust
4.6.1. Trust is not necessarily symetrical
4.6.2. True transparency involves giving people information about the connection between a reviewer and the reviewed, such as a list of all published reviews and comments by and about a user.
4.6.3. Yelp and false reviews
4.6.3.1. A one star increase in Yelp ratings garnered a 5-9% increase in revenues for a business.
4.6.3.2. Meiting-Dianping, a chinese mashup of Yelp and Groupon boast over 200 m monthly users. When redeemed the sites's coupons help to confirm that the reveiwer actually set foot on the premises and purchased the service or product being reviewed.
4.6.4. Social media will increasingly be utilized in the evaluation of people's trustworthiness before and during their interactions with many institutions.
4.6.4.1. AllState Insurance covers 10% of American households hypothesized that its customers were more likely to file a false claim if people in their social network had themselves previously filed false claims in the past.
4.6.4.2. Rapleaf had a massive collection of email addresses and social netwokr data, with most of the FB data bought from apps that accessed people's FB acconts with their permissions.
4.6.4.3. Rapleaf provided info of connections between people. The db allowed Allstate to identify customers whose friends were also Allstate customers, and Allstate then could determine at what level to investigate a client's claim based on his friends' history with the insurer.
4.6.4.4. German Insurance Friendsurance, peer to peer insurance. To setup an account, you need tow or more people to indicate that they will contribute a specified amount in support of a deductible.
4.6.4.4.1. In a way, customers vouch for the honesty of their claims to their friends, and their friends vouch with their wallets for the veracity of the claim when the insurer has to pay the amount due.
4.7. The Large Context
4.7.1. We take counsel from our friends and families and notice what they consider to be appropriate and inappropriate.
4.7.1.1. FB also links people who share a computer or mobile device, even on a single occasion
4.7.1.2. Biocatch authenitcates a user through a usage fingerprint based on keystroke and mouse movements patterns.
4.7.2. Decisions are always made in context
4.7.2.1. our physical environment also influence the decisions we make.
4.7.2.2. We make different decisions based on the time of day, the weather outside, level of fatigue or happiness among others.
5. Chp4: Context and Conditions
5.1. The case of jeff grey and publicly video recording police action
5.1.1. Not all govt's give its citizens the right to access the type of records that helped Gray fight his arrest.
5.1.2. The complications of continuous recording can be seen in the Orlando police case. A driver successfully challenged a charge DUI when the dashcam video wasnt available.
5.2. A personalized Point of View
5.2.1. People are uncomfortable with data power assymetries.(they have data, you dont)
5.2.1.1. fear of information imbalance
5.2.1.2. fear of dissemination
5.2.1.3. fear of permanence
5.3. From Where You are to Who You're with
5.3.1. Todd Humphreys, Univ of Texas At Austin professor argues that within a decade we will attach miniature geolocation trackers to nearly everything we own.
5.3.1.1. The rise of privacy preserving tech
5.3.1.1.1. gps jammers
5.3.1.1.2. gsm jammers
5.3.1.2. Remember your phone switches from celltower to celltower as you travel, so even without GPS, your location can still be tracked.
5.3.1.2.1. Photos shared online have GPS in their metadata.
5.3.1.3. another source of data is from pictures of you taken by OTHER people. The picture's metadata contains location information .
5.3.2. Tencent's QQ and identify verification
5.3.2.1. noticed a lot of prostitutes masquering in QQ
5.3.2.2. Using spoofed Photos.
5.3.2.2.1. Tencent then developed a program fro dynamically verifying a profile using video.
5.3.2.2.2. An app would direct a person to a community manager who would ask the user to take various poses like touch your right ear, shrug your left shoulder etc. if the face on the video matched the uploaded photos, the profile was then verified.
5.3.2.2.3. but there were still thousands so AI is now used
5.3.3. WorldPay developed a PIN camera for identification
5.3.3.1. Pin entry Device Camera
5.3.3.1.1. captures pictures as you enter your PIN in ATM machines
5.3.3.2. at point of sale, capture facial photos
5.3.3.3. central database of these biometric data
5.3.3.4. if the face doesnt match, clerks will be prompted to ask for another ID before the sale can go through
5.3.4. MasterCard allows users to setup a biometric profile.
5.3.4.1. upload a scan of fingerpooint
5.3.4.2. at point of sale, submit a fingerprint and selfie video
5.3.4.3. users are asked to blink (to prevent mask)
5.3.5. Iris Pattern recognition
5.3.5.1. Carnegie Mellon researchers were able to capture IRIS info from the sidemirror of a car.
5.3.5.2. In India, iris scans of 1 billion people are taken as part of National ID card
5.3.5.3. camera pictures up to 10m away! and iris do not degrade with age, unlike fingerprints
5.3.6. Vigilant Solutions
5.3.6.1. after your mobile phone, the car is next proxy for your location
5.3.6.1.1. uses license plate information from network of cameras and optical character recognition to identify every number plate.
5.3.6.1.2. Claims to add about 100m license plates each month
5.3.6.2. Two ways to query Vigilant
5.3.6.2.1. search for a license plate and learn when and where it had been
5.3.6.2.2. search for license plates at the location of a crime
5.3.6.2.3. your car is your proxy for location
5.3.6.3. users
5.3.6.3.1. Police
5.3.6.3.2. car companies looking to repossess the cars
5.3.7. Ambient Sounds
5.3.7.1. Inside the home
5.3.7.1.1. Devices
5.3.7.2. Outside the home
5.3.7.2.1. Soundscape communicates physical location as well
5.4. Baring Your Heart
5.4.1. London Based Realeyes
5.4.1.1. used sensors to play targeted videos based on emotions displayed by viewers. LG in men's toilet.
5.4.2. Paul Ekman, Univ of California measure the physiological effects of 6 basic emotions: anger, sadness, fear, contempt, surprise and happiness.
5.4.2.1. Facial Action Coding System (FACS)
5.4.2.2. Physicological indicators are universal
5.4.2.2.1. micro expressions
5.4.2.3. Adviser to San Diego company Emotient
5.4.2.3.1. smile detector, used in sony digital cameras to take pictures as soon as people start smiling
5.4.2.3.2. adapting it for use in Hospitals to detect pain in patients
5.4.2.3.3. emotient was acquired by Apple
5.4.2.4. Affectiva: measure emotional response to ads and televised debates
5.4.2.4.1. training data from call center (lablled)
5.4.3. Live Ops and Mattershight
5.4.3.1. uses voice detection software to match representatives with customers
5.4.3.2. strong regional accents from callers are matched to an agent with the same features
5.4.3.3. detect when to escalate to supervisors
5.4.3.3.1. outgoing, serious, shy personality types ?
5.4.3.4. IBM baby cries , tone and volume of the cry to help parents
5.4.3.5. infrared based heart rate
5.4.3.6. Blood test can identify emotions related to fear, etc.
5.4.4. Activity Trackers like fitbit, withings and garmin vivo series
5.4.5. US Defense dept's Fearbit
5.4.5.1. sensors that can smell chemicals
5.4.5.2. based on graphene
5.4.5.3. Stress might be detected from a person's breath
5.4.5.4. Steering wheels could analyze a persons sweat for emotional levels
5.4.6. Othello's error
5.4.6.1. Othello falsely presumed fear from his wife signalled infidelity
5.4.6.2. correctly detecting the emotion, but not the cause of the emotion
5.5. Where your focus is
5.5.1. easy to identify the general direction and duration of your gaze in real time using video camera.
5.5.1.1. pupils dilate when taking in new information
5.5.2. Difference between amateurs and professionals based on eye movements
5.5.2.1. Toby's eye tracking glasses
5.5.2.2. can be attached to a computer monitor
5.5.2.3. needs to be calibrated
5.5.2.3.1. can be merged with ecg etc, combining multiple streams of phyhsiological data
5.5.3. eye tracking can help doctors give better diagnosis
5.5.3.1. gaze control system
5.5.3.2. patient monityoring
5.5.3.3. detects when a person is about to 'zone' out based on the pattern of eye drifting
5.5.4. Harvard Study by Daniel Gilbert
5.5.4.1. people reported being zoned out about 20% of the time.
5.5.4.2. eye trackers can be used to compensate people based on their attention
5.5.4.3. eye trackers can detect
5.5.4.4. People report being unhappy when their minds wandered.
5.5.5. sociometric badges that monitor who you are with
5.5.5.1. contains: accelerometer, infrared LED and sensor to infer whom the person is facing.
5.5.5.2. Also contains bluetooth to capture proximity to other badges.
5.5.5.3. construct sociometric graphs, who sits next to whom etc.
5.5.5.4. indicates engagement
5.5.5.4.1. including speech tone etc.
5.5.5.5. reveals team cohesion
5.5.6. Ben Waver, study for Bank of America
5.5.6.1. Call center employees performing better when they have more face to face meetings in and outside of work
5.5.6.2. staff tenure etc.
5.5.6.3. lower stress levels
5.5.6.4. more simultaneous breaks, more opportunities for interactions
5.5.6.4.1. team performance increased by 25%
5.5.7. FMRI studies
5.5.7.1. allows scientist to see what part of the brain are activated.
5.5.7.2. wireless infrared imaging -- portable
5.6. Witness for the People
5.6.1. Zilla Van Den Born Fake Vacations
5.6.1.1. Fake photos of pretend vacation
5.6.1.2. easy to manipulate realities
5.6.2. More and more data will make it hard to fake it
5.6.3. these are the honest signals.
6. Chp5: Seeing the Controls
6.1. What can you demand to see about your data?
6.1.1. instead of controlling at point of creation
6.1.2. concentrate on the data refinery side
6.1.2.1. tool to understand how they arrive at the conclusions
6.1.3. your power lies in choosing those refineries which offer tools that increase transparency and agency
6.2. the suggested framework
6.2.1. Two rights to increase data refineries transparency
6.2.1.1. the right to access your own data
6.2.1.1.1. Lets you see and interpret what data is collected by and about you.
6.2.1.1.2. Must also include social graph data - you and others
6.2.1.1.3. ability to see your data in context
6.2.1.1.4. Facebook facial recognition was turned off in Europe for privacy reasons
6.2.1.1.5. Who really owns the data?
6.2.1.2. the right to inspect data refineries comprising of:
6.2.1.2.1. You need to see data about the data refinery!
6.2.1.2.2. three measures for inspecting a refinery
6.2.1.3. In Plain sight
6.2.1.3.1. a dashboard for easy viewing to display the 3 metrics above
6.2.1.3.2. distill and communicate the 3 metrics
6.2.1.3.3. Option for users to get notifications when the ratings change.
6.2.2. Four rights to increase AGENCY
6.2.2.1. right to amend data
6.2.2.2. right to blur your data
6.2.2.3. right to experiment with the refineries
6.2.2.4. right to port your data
7. Chp6: Taking the Controls
7.1. Agency for the People.
7.2. The right to amend data- free expression
7.2.1. sumerian priest that recorded the data on clay tablet. Maintaining the data was a way to maintain power.
7.2.2. Pete Warden, co founder of Jetpac says that today we're dealing with yet another period of overzealous protection of information.
7.2.2.1. Manual data verification does not scale. Take advantage of Machine learning for this.
7.2.3. instead of policing the data, give the USERs the ability to make their own mark on the record.
7.2.3.1. right to amend is better than right to be forgotten
7.2.4. right to amend co-owned data?
7.2.5. People are motivated to amend a record when they are likely to benefit from the change.
7.2.5.1. a homeowner may want to correct a wrong assessment on their property
7.2.5.2. the right to amend is sp important when data might harm you .
7.2.6. tie amendment to a person
7.2.6.1. Metadata of an amendment can be used to validate
7.2.6.2. Power Utility has its own distinct frequency or measure.
7.2.6.2.1. can be used to verify location
7.2.6.3. use of blockchain
7.2.6.3.1. every change is transparent.
7.2.6.3.2. consortium gets full read and write
7.2.6.3.3. public blockchain is better.
7.2.7. refineries must commit resources to support the right of users to amend data.
7.3. The right to blur your data- self determination
7.3.1. gives users the power to determine the level of detail that can be shared.
7.3.1.1. GPS or beacon location
7.3.1.2. Eric Horvitz, Microsoft Research: Setting spatial resolution
7.3.1.2.1. granularity settings is a function of your needs
7.3.1.2.2. based on variables like time of day etc.
7.3.1.3. personal characteristics like age, sex, weight, height, ethnicity, religion, employer, industry and occupation
7.3.1.3.1. Linkedin case.
7.3.1.3.2. blurring product data
7.3.1.3.3. blurring data at the source is irreversible.
7.3.1.3.4. so blurring is best LATER.
7.3.1.4. try introducing random noise to the data.
7.4. The right to experiment - exploration
7.4.1. letting users play with the possibilities
7.4.1.1. like how search results are presented. give users the right to experiment with the settings so the results can be ordered some other way
7.4.1.1.1. options
7.4.1.2. you can build a mental model of how the data refinery works.
7.4.1.3. hipmunk case
7.4.1.3.1. flight sorted by 'agony'
7.4.2. however, exposing these knobs may be risking the data refinery's trade secrets
7.4.3. might expose that the interests of the users and data refineries are not aligned.
7.4.4. hueristics
7.4.4.1. availability
7.4.4.2. representativeness
7.4.4.3. anchoring
7.5. The right to port your data- focused on increasing agency
7.5.1. need to use your data freely to whatever scope you like and with any recipient of our choosing.
7.5.2. data continue to exist where they were ported from.
7.5.3. Case of student transcripts
7.5.3.1. validation by destination of the source
7.5.4. Reputational Data - Uber drivers
7.5.4.1. ensures that reputation travels with people, just as it does in the physical world.
7.5.5. Data trump algorithms
7.5.5.1. data must flow in both directions
7.6. Influencing the Machines
7.6.1. let humans do what they are good at and let machines do what they are good at
7.6.1.1. case of Bosch ABS on mercedes and BMW
7.6.1.2. 1978. Safety test shows ABS is more superior
7.6.1.3. ABS+ Human Combined made it safer
7.6.1.3.1. future: AEB automatic emergency braking
7.6.1.3.2. reduces accidents by about 50%
7.6.1.4. Adaptive cruise control
8. Chp7: Rights into Realities- applying the power of transparency and Agency
8.1. buying on your own terms
8.1.1. Which type of data most helped purchase decision
8.1.1.1. browsing data?
8.1.1.1.1. "Customers who view this item also viewed...?
8.1.1.2. purchasing data?
8.1.1.2.1. "Customers who bought this item also bought
8.1.1.3. combination data
8.1.1.3.1. "Customers who viewed this item eventually bought.."
8.1.1.4. social data proves influential in amazon when people found the relationship between clicks and the eventual purchase of others (customers who viewed this item eventually bought...")
8.1.2. NZ based Icebreaker assigned a unique alphanumeric string.
8.1.2.1. it also allowed Icebreaker to track where fakes were showing up around the globe!
8.1.3. Food can also be tracked. QR codes to confirm authenticity
8.1.3.1. traceable products
8.1.4. Identifier data can be joined with sensor data
8.1.4.1. applied DNA
8.1.5. MIT's trash tags
8.1.5.1. understand the rate of recycling in the community
8.1.6. Air Travel example
8.1.6.1. no need to freeze the terms of service
8.1.6.2. amend the ticket with flexibility if you can be given cash back
8.1.6.3. match and keep more revenue
8.2. The future of finance
8.2.1. Pay with Affirm
8.2.1.1. short term payment company
8.2.1.2. reinvent consumer credit using social data to give people access to credit.
8.2.1.2.1. using web browsing behavior, frequency of mobile phone calls and text messages and even the operating system of the mobile phone.
8.2.1.2.2. check if the applicant is active in an online community such as GitHub.
8.2.1.3. affirm will ask for temporary read access to a checking account in order to analyze purchasing and income patterns.
8.2.2. Upstart provides credit to people in their 20s and 30s. to refinance credit card debt
8.2.2.1. uses the university you attended
8.2.2.2. your major
8.2.2.3. your classes and the grades you got and your SAT to predict salary growth over the next several years and figure out your repayment likelihood
8.2.3. Sesame Credit
8.2.3.1. using Alibaba's wealth of transaction and communication data from Alipay.
8.2.3.1.1. 14b on singles day 11.11.11
8.2.3.2. when people go dutch, alipay provides that payment option on the bill. This gives alibaba real world information not only about where oeple are eating but also with whome they are eating to calculate sesame scores!
8.2.3.2.1. know who they are eating with
8.2.4. semantic analysis of your tweet
8.2.4.1. allstate flags your connections.
8.2.5. social unfriending with people of low credit scores.
8.3. Fair Employment agency
8.3.1. JP Rangaswami allowed his email to be read by his direct reports.
8.3.1.1. employee bickering was taking up most of his time.
8.3.1.1.1. he allowed employees to view both inbox and outbox
8.3.1.1.2. resulted in drop of complaints.
8.3.1.2. he discovered that people were much more interested n his outbox than his inbox.
8.3.1.2.1. What he was saying was more important than what others were telling him.
8.3.2. Linkedin Sunday
8.3.2.1. noticed a lot of activity on the website on Sept 14, 2008
8.3.2.1.1. turns out a lot of Lehman Brothers employees were frantically reaching out to connections updating resumes and downloaidng contact information.
8.3.2.1.2. Linkedin suspected that the bailout plan at Lehman failed.
8.3.2.1.3. Employee exit is a bad sign
8.3.3. Social data can be used to optimize who they work and when they work
8.3.3.1. Greg Tanaka, founder developed models to generate predictions of store traffic and determine how many staff are needed to attend to customers. Optimizing staffing levels is key
8.3.3.2. inter personal dynamics may affect morale
8.3.3.3. Percolata model
8.3.3.3.1. camera and microphone
8.3.3.3.2. measure level of engagement from sensing the overall noise level
8.3.3.3.3. identify popular aisles
8.3.3.3.4. increased revenues while decreasing staffing costs
8.3.3.4. increased agency improves the entire ecosystem
8.3.3.5. Used for locating talent
8.3.3.5.1. Quora and MoData
8.3.3.5.2. Upwork
8.3.3.5.3. freelancers
8.3.3.5.4. ask for the right to port the ratings
8.3.4. Learning on the Playground
8.3.4.1. Schools kill creativity
8.3.4.2. classroom arrangements needs to be changed
8.3.4.3. testing based on student's mastery of fact.
8.3.4.4. teaching students how to ask good questions!
8.3.4.4.1. you can forget facts, but not the understanding
8.3.4.5. Learning Catalytics example
8.3.5. Precisely what the Data Ordered
8.3.5.1. Xray and its effects on the medical compared to 1b smartphones and 100m activity trackers
8.3.5.1.1. MRI allows soft tissue to be seen as well
8.3.5.1.2. a persons whole genome is now available
8.3.5.2. OpenNotes by Dr Tom Delbanco of Harvard Medical shared patients records with doctors and allowed patients to access them.
8.3.5.2.1. pilot study with 19,000 patients
8.3.5.2.2. patients are notified when a doctor added a note
8.3.5.2.3. patients asks for corrections
8.3.5.3. Vitality
8.3.5.3.1. discounts when shopping cart data shows healthy foods
8.3.5.4. cluster analysis for health
8.3.6. A fair deal?
8.3.6.1. MonkeyParking case of bidding out a parking space.
8.3.6.2. matching drivers to available parking space
9. Epilogue: Into the Sunlight
9.1. Allegory of the Cave
9.1.1. The key life lesson from Plato’s Allegory of the Cave is to question every assumption you have about the reality you call “real.”
9.1.2. They are shadows, not reality
9.1.3. it takes time for us to adjust to the 'light' and seeing the real world.
9.1.4. We must be free to see, and free to act