What is the data revolution that leveraged the big data that led the president to victory in the presidential election?
Psychologist Mikaru Kozinski has devised a method for finely analyzing the nature of the user and what he likes from "What did the Facebook user" like "?" Mr. Kozinsky's research results seem to have played a part in Mr. Donald Trump's presidential election victory.
The Data That Turned the World Upside Down - Motherboard
At 8:30 on November 9, 2017, Mr. Kozinsky woke up at the Sunne's Hotel in Zurich, Switzerland. Mr. Kozinsky was a 34 - year - old psychologist and came to Switzerland to attend the Swiss Institute of Technology in order to lecture on the dangers involved in big data and digital revolution. Mr. Kozinsky is doing a similar lecture all over the world and is a person known as a leading expert in psychology using psychometric measurements and data. When I woke up from the bed and turned on the TV, Mr. Trump seemed to have won the presidential election contrary to most predictions, he said he was looking at news reports such as victory celebrations and voting conditions in each state. And it seems that Mr. Kozinsky felt that the result of this election had some connection with his own research.
At the same time as this, a minor data analysis company based in London based in the UKCambridge AnalyticaHad released one press release. Among them are comments of Alegsander Jakob Ashburner Nicks, "We are thrilled that the innovative approach using our data played an important role in the victory of the next trump president" It was quoted.
Cambridge Analytica led by Nicks CEO played an important role in online campaign for Mr. Trump 's presidential election,British withdrawal from the European Union(Breguzzi) also has a track record of playing a major role. Mr. Kozinsky made his digital revolution possible from his own research, Mr. Nix executed his know-how in the real world, President Trump got a big advantage there.
Big data that played a major role in the presidential election refers to information that we have digitally tracked every action we do online or offline. It is a big data collection of all kinds of information such as "what you purchased with the card" by the Internet user, "what you entered on Google, what you searched for" "what you checked on your smartphone". In particular, what you like is an important factor, for example, Google's search for "lowering blood pressure", followed by "advertisement of hypertension treatment" advertisement is displayed, etc. Study on "what the user likes" It can be said that the results that have been reflected are reflected.
And Cambridge Analytica, which supported President Trump 's online campaign, revealed that the range of utilization of big data is wider. It seems that this is starting from Kozinsky's psychometric center in Cambridge University in 2014. "Psychometric measurement" is also called "psycho graphics" and is a research field that focuses on measuring the psychological characteristics of individuals. In the 1980's, "Big Five" which a research team by a psychologist tried to evaluate psychological state on the basis of five personality characteristics of a person is "psychometric" running. The five personality traits are "openness (how long is it open for new experiences?)" "Seriousness (how long is it a perfectionist)?" "Extroversion (how much sociable?)" "Comfort How prudent and cooperative? "" Neurosis (easily distracting?) "Was. "Big Five" is also known as "OCEAN", it means that we can evaluate humans relatively accurately.
By using this evaluation, it seems that people understand "what you are seeking", "what is afraid?", And how those feelings affect each person. Because of its usefulness, "Big Five" has become a standard technique in psychometry. However, this approach was thought to have problems with data collection. Because it was necessary to answer complicated and very personal questionnaire to accurately perform psychometric measurement using "Big Five".
However, after that, the Internet and Facebook were born. Kozinsky developed an application called "MyPersonality" using information on Facebook. This application makes it easier to measure psyche simply by answering quite a few questionnaires compared to "Big Five". When answering questions in the application, the user is able to receive what is called "personality profile", and it seems that some of the options included "using Facebook profile". Kozinsky succeeded in obtaining a huge dataset called a Facebook profile by adding the option of "using a Facebook profile".
Likewise, taking a quiz questionnaire on-line over several years and calculating respondent's "Big Five" from that answer. And from estimating postings to Facebook, we estimate everything such as sex, age, address, etc. from "What do you like like?" And increase its accuracy. As a result, for example, "Men who likes" MAC "of cosmetics are many homosexuals" and "heterosexuals like Wu Tang Clan" "People who like Lady Gaga are diplomatic" "I like philosophy People are introverted "and it seems that it seems to become visible.
Mr. Kozinsky's research made it possible to generate reliable forecasts if we can collect and combine dozens, hundreds and thousands of different data points, even though the individual information is very small. And in 2012, Kozinsky's research says, "By analyzing the average of 68 Facebook's 'Like', the user's skin color (95% correctness rate), sexual preference (88% correct answer) , It is possible to hit either the Democratic Party or the Republican Party (correct answer rate 85%) ". Using this method, it is possible to estimate whether alcohol, tobacco, drugs are loved, what kind of religion they join, and who's divorced.
Even then, Mr. Kozinsky refined his application, and eventually ranked from ten "goods" to average colleagues, 70 "friend" friends, 150 "good" parents , 300 hits and tastes can be estimated more accurately than the partner with "good likes". When publishing this research result, Mr. Kozinsky said he received "threat of lawsuit" and "offer of work" by phone, both of which were from Facebook.
You can estimate fairly detailed personal information from "likes" on Facebook, but if you have a built-in motion sensor on your home phone, for example, it is perfect for estimating "how long have you been on a trip" It may be possible to obtain information. Mr. Kozinsky said, "Using smartphones is like filling up huge psychology questionnaires," and that's why smartphones are so packed with personal information.
In essence, "Mr. Kozinsky invented" human search engine "," Motherboard of international news media. Using this, it was possible to bare up to "worried (someone) father", "angry introverted person", even "indecisive Democratic supporters", I was not aware of that Mr. Kozinsky, I noticed that, it was Cambridge Analytica that applied to the election campaign.
"Most of the messages by President Trump is based on data," says Nicks of Cambridge Analytica. President Trump and Mr. Hillary · Clinton launched a verbal battleThe 3 rd session of the 2016 US presidential electionSo, the trump camp advertised 175,000 different variations on Facebook to find the answer as to what kind of discussion should be held. Advertisement seems to have been different little by little from headline, color, caption, photo, movie and so on.
Also, for example, the trump camp against the Haitian district in Miami City "The Clinton Foundation had donated a relief donation for Haiti earthquake"We delivered the news that produced people like Haiti who do not want to vote for the Hillary camp. Nix said, "We can talk to the target people on a village or apartment complex basis", and we are managing the election campaign by delivering advertisements and news to more specific areas and specific targets It is clear that you know.
Also, it was aimed at keeping people from the left wing, African-American, young women, etc. potentially voting for the Clinton campaign away from the election campaign,BloombergI am telling you. The "dark post" posted by the trump camp to lower Clinton's reputation is a newsfeed style advertisementFacebook advertisementIt is delivered to the people of the American people using, for example, "Hillary · Clinton calls a black man as predator".
Although it won the splendid campaign by the digital unit of the trump camp, it did not see that identity during the election campaign. Because the trump camp digital unit does not target television which is the mainstream of media, because it used social media and digital TV instead.
Cambridge Analytica, who has been struggling with the digital strategy at the trump camp, is a team of only about twelve, but in the month of July 2016 it is 100,000 dollars (about 11 million yen), in August it is 250 thousand dollars (about 28 million yen), September According to Nicks CEO, Cambridge Analytica earned over $ 15 million (about 1.7 billion yen) through the presidential election.
Also, in July of 2016, exclusive applications were dealt to each trainee camped door visitor. The application seems to indicate what kind of political opinion the resident of each person visits in such a personality type. The same application was also used during Breguzz's exercise. The application discriminated residents who could accept opinions of the trump camp, and visitors were able to conduct house conversations effectively by advancing conversation according to the guidelines prepared by the character types of the residents That's right.
During the US presidential election, Cambridge Analytica classified American citizens into 32 personality types, especially focusing on 17 states. As Mr. Kozinsky discovered from "Big Data" "There are many homosexuals in men who like the MAC of cosmetics", Cambridge Analytica says, "From the preference of American-made cars, they are ally of potential trump camps "I discovered. It is also based on these data that the trump camp was to focus on Michigan and Wisconsin.
A similar approach was also taken by Hillary camp but it is unknown whether or not profiling with psychometric measurement was done.
in Note, Posted by logu_ii