How to Interpret a Vector Set

A Whit.li vector set contains information that describes the various traits of a person.   Each Whit.li vector set contains multiple vectors that have detailed information about an individual based on the individuals psycho-social profile.  The following examples show how to decode a Whit.li vector set.  

!

Example Vector Set: KEY 1

K1:V1(2,30.26715,-97.74306,790390,34);V2(1.53,17.62,3.22,1,3.68);V3(6.58,6.4);V4(5,9,24,2,21,6,0,6,18)

Vector 1

V1 contains demographics such as: sex, the individual's name, schools attended, interests, and likes.  

V1 (2,30.26715,-97.74306,790390,34) 

In this example, the individual is:

• female
• their current location latitude is 30.26715
• their current location longitude is -97.74306
• their current location population is 790390
• this person is 34 years old

Vector 2

V2 contains raw personality scores that include: extraversion, agreeableness, neuroticism, conscientiousness, openness. These scores are based on the Big 5 personality traits. More information on the research behind the Big 5 can be found at http://en.wikipedia.org/wiki/Big_Five_personality_traits. Below is a brief description of each of the traits.

V2 (1.53,17.62,3.22,1,3.68) 

• extraversion = 1.53
• agreeableness = 17.62
• stability = 3.22
• conscientiousness = 1
• openness = 3.68

• Extraversion – (outgoing vs. reserved). People who are high in extraversion are very outgoing and get energy from engaging with other people. People who are low in extraversion are more reserved and get energy from personal time.  Minimun score is 0, Maximum Score is 100, Mean (in our data) is 1.37
• Agreeableness – (flexible and warm vs. directed and businesslike). People who are high in agreeableness tend to be flexible and cooperative where as people who have lower agreeableness tend to be more businesslike and opinionated.   Minimun score is 0, Maximum Score is 100, Mean (in our data) is 16.24
• Stability – (steady vs. sensitive). People who have high stability are steady and unflappable, even in situations that are stressful. People who have lower stability may respond emotionally in stressful situations. Minimun score is 0, Maximum Score is 100, Mean (in our data) is 2.36
• Conscientiousness – (efficient vs. carefree). People who have high conscientiousness act dutifully and precisely. People who are low in Conscientiousness are carefree and spontaneous. Minimun score is 0, Maximum Score is 100, Mean (in our data) is 1.55
• Openness to experience – (curious vs. consistent). People who have high openness to experience are curious and like a variety of experience. People who have low openness to experience prefer consistent experiences and value known traditions.  Minimun score is 0, Maximum Score is 100, Mean (in our data) is 2.07

Vector 3

V3 contains raw sentiment scores that include overall sentiment and sentiment for the past week (lowest sentiment = 1, highest sentiment = 9). Higher sentiment is positive representing a more positive attitude. Lower sentiment is negative representing a more negative attitude. Sentiment is analyzed using Natural Language Processing. More information on the methodology used can be found at: http://a.parsons.edu/~spani621/thesis/context/ANEW.pdf

Vector 3 is composed of two components:

 

• component 200-Overall Sentiment
• component 201-Recent Sentiment

 

Example: V3(5.96,6.89)

 

• overall sentiment = 5.96
• sentiment for the past week = 6.89

Vector 4

V4 contains interest categories.  For interest categories, the number displayed is a count.  For example, this individual list 5 TV shows or movies that are their favorites.

V4 (5,9,24,2,21,6,0,6,18)

• TV/movies = 5
• food = 9
• going out = 24
• sports = 2
• volunteer = 21
• technology = 6
• books = 0
• music = 6
• products = 18

!

Example Vector Set: KEY 2

K2: V5(3,2.17,18.5,1.73,2.82,2.6,3,2,3,1,1,2,3,12)

Vector 5

V5 contains components necessary for comparisons using contexts 100-104.

V5 (3,2.17,18.5,1.73,2.82,2.6,3,2,3,1,1,2,3,12,6.58) 

In this example, the comparison shows:

• sentiment is high
• extraversion raw score is 2.17
• agreeableness raw score is 18.5
• conscientiousness raw score is 1.73
• neurosis raw score is 2.82
• openness raw score is 2.6
• extraversion is high
• agreeableness is average
• neurosis is high
• conscientiousness is low
• openness is low
• sex is female
• age group is 25-34
• relationship status is default (or not specified)
• sentiment raw score is 6.58

Detailed Vector Breakdown

• 1. Sentiment L/A/H (Low = 1)
• 2. Extraversion Raw Score
• 3. Agreeableness Raw Score
• 4. Conscientiousness Raw Score
• 5. Neurosis Raw Score
• 6. Openness Raw Score
• 7. Extraversion L/A/H (Low = 1)
• 8. Agreeableness L/A/H (Low = 1)
• 9. Neurosis L/A/H (Low = 1)
• 10. Conscientiousness L/A/H (Low = 1)
• 11. Openness L/A/H (Low = 1)
• 12. Sex (Male=1, Female=2)
• 13. Age Group

1. 1 = 0-17
2. 2 = 18-24
3. 3 = 25-34
4. 4 = 35-44
5. 5 = 45-54
6. 6 = 55-64
7. 7 = 65+

• 14. Relationship Status

1. 1 = Single
2. 2 = It's complicated
3. 3 = Divorced
4. 4 = Separated
5. 5 = In an open relationship
6. 6 = Widowed
7. 7 = In a relationship
8. 8 = Engaged
9. 9 = Married
10. 10 = In a civil union
11. 11 = In a domestic partnership
12. 12 = default (or not specified)

• 15. Sentiment Raw Score

!

Example Vector Set: KEY 6

This key returns derived internal values like sentiment, personality and “truthiness”. 

K6:V10(-0.09,0.27,0.2,-0.59,0.03);V3(5.96,6.89);V11(9,0)

Vector 10

V10 contains additional personality score information.  The vector returns the amount that a personality score is above or below the mean score for that component in terms of standard deviation.  A result close to 0 indicates average, negative indicates below average, and positive indicates above average. There are five components that comprise this vector. 

• component 150-Extraversion (standard deviations above or below Whit.li’s mean)
• component 151-Agreeableness (standard deviations above or below Whit.li’s mean)
• component 152-Neurosis (standard deviations above or below Whit.li’s mean)
• component 153-Conscientiousness (standard deviations above or below Whit.li’s mean)
• component 154-Openness (standard deviations above or below Whit.li’s mean)

Example: V10(-0.09,0.27,0.2,-0.59,0.03)

In this example, the individual is:

• .09 standard deviations below the mean for Extraversion
• .27 standard deviations above the mean for Agreeableness
• .2 standard deviations above the mean for Neurotic (the opposite of stability)
• .59 standard deviations below the mean for Conscientiousness
• .03 standard deviations above the mean for Openness 

Vector 3

Vector 3 contains raw sentiment scores that include overall sentiment and sentiment for the past week (lowest sentiment = 1, highest sentiment = 9). Higher sentiment is positive representing a more positive attitude. Lower sentiment is negative representing a more negative attitude. Sentiment is analyzed using Natural Language Processing. More information on the methodology used can be found at: http://a.parsons.edu/~spani621/thesis/context/ANEW.pdf

Vector 3 is composed of two components:

• component 200-Overall Sentiment
• component 201-Recent Sentiment

Example: V3(5.96,6.89)

• overall sentiment = 5.96
• sentiment for the past week = 6.89

 Vector 11:

Vector 11 provides a mapping of the personality to one of 42 personality types based on the users most extreme personality traits and a way to assess the relative truthfulness of the poster.  The personalities are values 1 - 40, decoded below and the truthfulness is a standard deviation away from average truth. A result close to 0 indicates average, negative indicates below average (less truthful), and positive indicates above average (more truthful).  This is best for analyzing a single user over time. Vector 11 is composed of two components.

• component 180-Personality Type (A value from 1 - 40 - decoded below)
• component 500-Truthiness as a standard deviation from a mean of truth.

Example: V11(9,0)

• 9 = Industrious
• Truthiness = 0, or average truthful

Personality Types - Decoded

!

Context Breakdown

New!  All contexts return the users common interests with a commonality score so that unusual interests can be identified.  This functionality works best with Facebook data for users.

• Context_ID 100 - WORK

To determine how well you’d work with someone, Whit.li compares your relative levels of Conscientiousness and Agreeableness according to scientifically validated measures of work compatibility.

The response returns a score from 0 to 20. In general, low compatibility is (0-9.99), medium compatibility is (10-14.99), and high compatibility is (15+).  The likes returns the id, name, type, and commonality (which describes how rare it is).

  Example response:

  {"score":40,"code":"","common_likes":[{"id":"172109382825604","name":"Whit.li- Explore Minds Like Yours","type":"LOCAL BUSINESS","commonality":"0.3687"}

  In this case, the context score is 40 and both users share Whit.li as a common like.  Commonality right now is a score from slightly above 0 (0.0001) to infinity, with 1 as the median. The closer to 0, the least common the like is and the larger, the more common.  Commonality is based within the type of the like.

• Context_ID 101 - SHOPPING

To determine how well you’d shop with someone, Whit.li combines divined measures of class, culture, age and lifestyle stage with underlying consumer personality such as whether your are principle oriented or action oriented or whether you prefer variety or routine.

The response returns a score from 0 to 95. In general, low compatibility is (0-59.99), medium compatibility is (60-84.99), and high compatibility is (85+). The likes returns the id, name, type, and commonality (which describes how rare it is).

Example response:

{"score":40,"code":"","common_likes":[{"id":"172109382825604","name":"Whit.li- Explore Minds Like Yours","type":"LOCAL BUSINESS","commonality":"0.3687"}

In this case, the context score is 40 and both users share Whit.li as a common like.  Commonality right now is a score from slightly above 0 (0.0001) to infinity, with 1 as the median. The  closer to 0, the least common the like is and the larger, the more common.  Commonality is based within the type of the like.

• Context_ID 102 - TRAVEL

To determine how well you’d travel with someone, Whit.li divines your underlying travel personality like Adventurer or Sightseer, using your personality and travel history, then we assess compatibility based on a proprietary matching algorithm.

The response returns a score from 0 to 50. In general, low compatibility is (0-19.99), medium compatibility is (60-40), and high compatibility is (40.01+). The likes returns the id, name, type, and commonality (which describes how rare it is).

Example response:

{"score":40,"code":"","common_likes":[{"id":"172109382825604","name":"Whit.li- Explore Minds Like Yours","type":"LOCAL BUSINESS","commonality":"0.3687"}

In this case, the context score is 40 and both users share Whit.li as a common like.  Commonality right now is a score from slightly above 0 (0.0001) to infinity, with 1 as the median. The  closer to 0, the least common the like is and the larger, the more common.  Commonality is based within the type of the like.

• Context_ID 103 - ROOMMATES

To determine how well you’d click as roommates, Whit.li compares your relative levels of Extraversion, Conscientiousness and Emotional Stability according to scientifically validated measures of compatibility, and further enhances these measures with information that determines the compatibility of your personal values.

The response returns a score from 0 to 40. In general, low compatibility is (0-19.99), medium compatibility is (60-30), and high compatibility is (30.01+). The likes returns the id, name, type, and commonality (which describes how rare it is).

Example response:

{"score":30,"code":"","common_likes":[{"id":"172109382825604","name":"Whit.li- Explore Minds Like Yours","type":"LOCAL BUSINESS","commonality":"0.3687"}

In this case, the context score is 30 and both users share Whit.li as a common like.  Commonality right now is a score from slightly above 0 (0.0001) to infinity, with 1 as the median. The  closer to 0, the least common the like is and the larger, the more common.  Commonality is based within the type of the like.

• Context_ID 104 - FRIENDS

To determine how well you would get along as friends, Whit.li combines all of your underlying personality and demographic traits in a proprietary algorithm to assess compatibility for long-term friendship.

The response returns a score from 0 to infinity. In general, low compatibility is (0-30.99), medium compatibility is (31-69.99), and high compatibility is (70+). The likes returns the id, name, type, and commonality (which describes how rare it is).

Example response:

{"score":40,"code":"","common_likes":[{"id":"172109382825604","name":"Whit.li- Explore Minds Like Yours","type":"LOCAL BUSINESS","commonality":"0.3687"}

In this case, the context score is 40 and both users share Whit.li as a common like.  Commonality right now is a score from slightly above 0 (0.0001) to infinity, with 1 as the median. The  closer to 0, the least common the like is and the larger, the more common.  Commonality is based within the type of the like.

• Context_ID 106 - NETWORKING

To determine how well you would get along meeting at a conference, Whit.li combines certain personality elements, sentiment, and common interests (including its rarity) in a proprietary algorithm to assess compatibility for a first time meeting.

The response returns a score from 0 to infinity. In general, low compatibility is (0-30.99), medium compatibility is (31-69.99), and high compatibility is (70+). The likes returns the id, name, type, and commonality (which describes how rare it is).

Example response:

{"score":55,"code":"","common_likes":[{"id":"172109382825604","name":"Whit.li- Explore Minds Like Yours","type":"LOCAL BUSINESS","commonality":"0.3687"}

  In this case, the context score is 55 and both users share Whit.li as a common like.  Commonality right now is a score from slightly above 0 (0.0001) to infinity, with 1 as the median. The closer to 0, the least common the like is and the larger, the more common.  Commonality is based within the type of the like.