You need to sign in or sign up before continuing.
Take a photo of a barcode or cover
659 reviews for:
Everybody Lies: Big Data, New Data, and What the Internet Can Tell Us about Who We Really Are
Seth Stephens-Davidowitz
3.82 AVERAGE
659 reviews for:
Everybody Lies: Big Data, New Data, and What the Internet Can Tell Us about Who We Really Are
Seth Stephens-Davidowitz
3.82 AVERAGE
I would have rated this more highly if he would've left out the little snippets showing his political bias. It really started making me question whether his other assertions were founded more with other biases in mind or based on good research.
informative
fast-paced
A pretty decent book but felt like a BTEC freakonomics, even down to the style and tone of writing.
4.5 Stars. The information in this book presented both devastating (racism, abuse) and hilarious (porn proclivities) conclusions about who we really are, based in large part on data from unhindered google searches. I wonder whether data I provide with my goodreads account will be used for some profound conclusion about human behaviors. Very quick read, highly interesting.
This is a very easy book to read. The whole book is about big data with a lot of practical examples of how it's used to solve everyday problems. This book is not a primer on developing big data datasets but shows the reader how data scientists make use of the data and some of the insights they can get from big data.
This book deals with how big data can be used to learn about people at a particular place and time. I learnt a lot of things about people and society with a lot of accepted norms being shown to be false by data. Oh yes, people always think the best (or worse) of themselves and others but the data show a totally different picture. Hence the title, everyone lies...except to a search engine.
Besides all the good new research that can be done with big data, the author shows the limitations of using big data and the difficulty of using correlation for causation. The type of datasets is important. More data is not always the answer but the right kind of data is important.
It's a fun book and I highly recommend it, even if you are not a data scientist.
This book deals with how big data can be used to learn about people at a particular place and time. I learnt a lot of things about people and society with a lot of accepted norms being shown to be false by data. Oh yes, people always think the best (or worse) of themselves and others but the data show a totally different picture. Hence the title, everyone lies...except to a search engine.
Besides all the good new research that can be done with big data, the author shows the limitations of using big data and the difficulty of using correlation for causation. The type of datasets is important. More data is not always the answer but the right kind of data is important.
It's a fun book and I highly recommend it, even if you are not a data scientist.
I’m surprised I hadn’t heard of this book before coming across it at a used book sale recently, since it came out in 2017 and covers human behavior topics in which I’m interested, with a unique angle.
I learned numerous things from the book, including the story about predicting horse race success based on the size of internal organs, and the fact that Floridians apparently stock up on strawberry Pop Tarts before hurricanes.
Stephens-Davidowitz is a clear and engaging writer, in the tradition of [a:Steven Pinker|3915|Steven Pinker|https://images.gr-assets.com/authors/1235758085p2/3915.jpg], who wrote the foreword to this book, but with a slightly quirkier tone. The big picture view that he conveys about the use of not only large quantities of data but innovative sources of data is truly fascinating.
On the other hand, some questions remain in my mind: The author presents valid evidence that the data from Google searches really does correspond to patterns in the world. But why exactly do people search for things that are not even searchable questions such as “I hate cold weather” (p. 112)? Maybe a study should be done to explain why people don’t at least make an effort to phrase it in the form of a question? I understand that the data doesn’t lie, but maybe it’s only capturing a strange subset of the behavior, or maybe peoples’ habits will change over time.
Also, there’s an emphasis on searches for offensive or threatening terms, and also on sexual topics, mainly because this seems to be what “Everybody Lies” most about. But if that’s the case, can we really trust the data from people who are logged in to PornHub with presumably authentic accounts, having accurately provided their gender and other personal info? Or are more people just finding videos to watch for free? Then there’s the study where teachers were paid for attendance that concluded that “The results were remarkable.” (p. 209). I’m not sure it’s so remarkable that people do better work if you pay them more.
The commentary on companies’ excessive use of A/B testing to basically exploit their target audience is spot on, and although the author allows that big data also benefits consumers in some cases, the power to abuse it is clearly in the hands of those who can pay for it.
Overall, I enjoyed the book; I’d say it was 3 to 4 star, and I’ll round up for polished execution.
I learned numerous things from the book, including the story about predicting horse race success based on the size of internal organs, and the fact that Floridians apparently stock up on strawberry Pop Tarts before hurricanes.
Stephens-Davidowitz is a clear and engaging writer, in the tradition of [a:Steven Pinker|3915|Steven Pinker|https://images.gr-assets.com/authors/1235758085p2/3915.jpg], who wrote the foreword to this book, but with a slightly quirkier tone. The big picture view that he conveys about the use of not only large quantities of data but innovative sources of data is truly fascinating.
On the other hand, some questions remain in my mind: The author presents valid evidence that the data from Google searches really does correspond to patterns in the world. But why exactly do people search for things that are not even searchable questions such as “I hate cold weather” (p. 112)? Maybe a study should be done to explain why people don’t at least make an effort to phrase it in the form of a question? I understand that the data doesn’t lie, but maybe it’s only capturing a strange subset of the behavior, or maybe peoples’ habits will change over time.
Also, there’s an emphasis on searches for offensive or threatening terms, and also on sexual topics, mainly because this seems to be what “Everybody Lies” most about. But if that’s the case, can we really trust the data from people who are logged in to PornHub with presumably authentic accounts, having accurately provided their gender and other personal info? Or are more people just finding videos to watch for free? Then there’s the study where teachers were paid for attendance that concluded that “The results were remarkable.” (p. 209). I’m not sure it’s so remarkable that people do better work if you pay them more.
The commentary on companies’ excessive use of A/B testing to basically exploit their target audience is spot on, and although the author allows that big data also benefits consumers in some cases, the power to abuse it is clearly in the hands of those who can pay for it.
Overall, I enjoyed the book; I’d say it was 3 to 4 star, and I’ll round up for polished execution.
This is worthwhile read although findings if not the methodology will likely be outdated within 10 years. Stephens-Davidowitz tells a story for a general audience. It's pretty simple. We, in some ways and to an extent, reveal true things about ourselves in our Google queries.
Men and women ask Google about their physical and relationship insecurities. Men are often concerned with their sexual performance duration, their penis size, and whether their significant other is cheating. Women are often concerned by the smell of their vaginas and why their significant others aren't having sex with them.
Both sexes, at least in our current time period, are interested in incestual porn on PornHub and some women show an interest in rape fantasies on PornHub. Here I should mention that the author rarely uses absolute numbers, but reports rates and comparisons to the opposite sex. So, don't be overly concerned about the prevalence, but you should be aware of the existence. Especially because these are taboo subjects.
He finds that rates of out gay men seem lower than expected in certain conservative and rural parts of the country, but that the missing gay men reappear in search queries. The same goes for the supposedly non-existent gay men in Iran and Sochi, Russa. Given this data, Seth expects about 5% of men to be gay.
Kids ask Google about dealing with abusive parents, especially in the aftermath of the 2008 recession. Women in states with restricted access, ask Google about DIY abortions. Racists search for racist jokes or combine a racial category with a strong negative adjective.
Some of the story confirms our cultural understanding of how the world works. Some of our cultural stories are disconfirmed. These, of course, interest me the most.
He has some warnings against how to use big data, and this data in particular, for individuals, businesses, and governments. For example, governments should be extremely cautious about trying to use it to predict or prevent crimes by an individual for compelling statistical and ethical reasons. A good use would allocate resources to a geography area, such as gay support groups expanding their presence to certain areas or government child services to others where abuse search terms are increasing. I was fascinated that racist search terms spiked during and after a President Obama speech haranguing Americans to be more tolerant, but a later speech that simply told stories of non-white Americans as soldiers, doctors, teachers, engineers, etc was associated with many fewer racist searches.
Two obvious pitfalls to be aware of when using this data. First, these queries are not directly measuring the things we are interested in - they are proxies. So, there's risk of conflating what we can measure with the truth. Two, since predictability does not always lead to explanatory theory, we risk a host of misunderstandings, such as the cause-effect relationship or the turkey problem. The turkey problem is the Thanksgiving turkey who lives a great life and continues to belief it will live a great life until, unbeknownst to it, a certain holiday arrives on the calendar.
Men and women ask Google about their physical and relationship insecurities. Men are often concerned with their sexual performance duration, their penis size, and whether their significant other is cheating. Women are often concerned by the smell of their vaginas and why their significant others aren't having sex with them.
Both sexes, at least in our current time period, are interested in incestual porn on PornHub and some women show an interest in rape fantasies on PornHub. Here I should mention that the author rarely uses absolute numbers, but reports rates and comparisons to the opposite sex. So, don't be overly concerned about the prevalence, but you should be aware of the existence. Especially because these are taboo subjects.
He finds that rates of out gay men seem lower than expected in certain conservative and rural parts of the country, but that the missing gay men reappear in search queries. The same goes for the supposedly non-existent gay men in Iran and Sochi, Russa. Given this data, Seth expects about 5% of men to be gay.
Kids ask Google about dealing with abusive parents, especially in the aftermath of the 2008 recession. Women in states with restricted access, ask Google about DIY abortions. Racists search for racist jokes or combine a racial category with a strong negative adjective.
Some of the story confirms our cultural understanding of how the world works. Some of our cultural stories are disconfirmed. These, of course, interest me the most.
He has some warnings against how to use big data, and this data in particular, for individuals, businesses, and governments. For example, governments should be extremely cautious about trying to use it to predict or prevent crimes by an individual for compelling statistical and ethical reasons. A good use would allocate resources to a geography area, such as gay support groups expanding their presence to certain areas or government child services to others where abuse search terms are increasing. I was fascinated that racist search terms spiked during and after a President Obama speech haranguing Americans to be more tolerant, but a later speech that simply told stories of non-white Americans as soldiers, doctors, teachers, engineers, etc was associated with many fewer racist searches.
Two obvious pitfalls to be aware of when using this data. First, these queries are not directly measuring the things we are interested in - they are proxies. So, there's risk of conflating what we can measure with the truth. Two, since predictability does not always lead to explanatory theory, we risk a host of misunderstandings, such as the cause-effect relationship or the turkey problem. The turkey problem is the Thanksgiving turkey who lives a great life and continues to belief it will live a great life until, unbeknownst to it, a certain holiday arrives on the calendar.
This book is filled with interesting facts about our behaviors, but I believe it could have been summarized more effectively. It's highly likely that your experience will differ if you approach the book with a different goal in mind. However, in my experience, it was quite repetitive.
i am glad i've read this book but i thought it would open up my eyes more. i've already known a lot of information written there but still, it showed me the insides and statistics of google and big brother watching us. i've even become more conscious and careful about what i google and what i do in real life. tho a big con for me is the number of sex-assosiations. but i guess all people are animals, so they can assosiate all the hard information with something they are already familiar with?