Take a photo of a barcode or cover
662 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
662 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
In the conclusion he calls this a more in-depth version of Freakonomics with the depth provided by the additional data available now. That's a fair summation - if you like data analysis or behavioral economics, or just want to know more about what data science is, this is a good choice. He shows how he can pick up trends from Google searches and what really can't be found with the current data. He shows fascinating and weird little factoids found through this and points to the failures of surveys and self-reported information versus predictive powers of large amounts of data. He sums up very well how the collected information can be useful on a broad scale while not applicable on an individual one with data about searches for suicide and murder.
He makes a strong argument about how natural experiments and greater data availability will strengthen the social sciences, making them as rigorous as the hard sciences. I have to admit this was a slow read - I began it with an ebook version and made very little progress, but I eventually switched to the audiobook and while still slow I did manage to finish it. The tables and charts demonstrating the effectiveness of A/B testing and the results of some of the natural experiments were useful to refer back to.
Overall, I would recommend because the methods and tools he talks about are ubiquitous and becoming more so while not being well understood outside of the technology community. These tools are changing the world around us, not just in academic theory, but in every dealing you have with a large corporate entity. They're using these tools on you so you might as well know what they are.
He makes a strong argument about how natural experiments and greater data availability will strengthen the social sciences, making them as rigorous as the hard sciences. I have to admit this was a slow read - I began it with an ebook version and made very little progress, but I eventually switched to the audiobook and while still slow I did manage to finish it. The tables and charts demonstrating the effectiveness of A/B testing and the results of some of the natural experiments were useful to refer back to.
Overall, I would recommend because the methods and tools he talks about are ubiquitous and becoming more so while not being well understood outside of the technology community. These tools are changing the world around us, not just in academic theory, but in every dealing you have with a large corporate entity. They're using these tools on you so you might as well know what they are.
This book felt like a warm hug from my Economics undergraduate curriculum. The gist is that with the public availability of Google search data, aka the stuff we type into our computers from the privacy of our homes, reveals our true sentiments, unlike surveys or interviews. This books explores that with random examples, which read like Gladwell with statistical backing.
informative
reflective
medium-paced
adventurous
funny
informative
reflective
fast-paced
Overview:
Human thoughts are very complex, making many different ways to access them subject to a variety of limitations. Big data via the internet is very revealing about people. What people search for is information that can reveal what they are thinking about. Providing a different way to learn about people. There are four ways that big data adds value to research which are 1) new types of data, 2) honest data, 3) zoom on small subsets, and 4) generate causal experiments. The internet is a primary source of data that uses messy traces, unlike traditional sources of data such as surveys and questionnaires which are very neat. Big data has many limitations as well, but it can help in a variety of situations. A problem with big data is that as it enables better predictions about human behavior, the information can be misused by corporations and governments leading to various forms of nefarious discrimination activities.
Human thoughts are very complex, making many different ways to access them subject to a variety of limitations. Big data via the internet is very revealing about people. What people search for is information that can reveal what they are thinking about. Providing a different way to learn about people. There are four ways that big data adds value to research which are 1) new types of data, 2) honest data, 3) zoom on small subsets, and 4) generate causal experiments. The internet is a primary source of data that uses messy traces, unlike traditional sources of data such as surveys and questionnaires which are very neat. Big data has many limitations as well, but it can help in a variety of situations. A problem with big data is that as it enables better predictions about human behavior, the information can be misused by corporations and governments leading to various forms of nefarious discrimination activities.
People tend to be way more honest in their search terms rather than what they claim about themselves. Depending on the context, they can be honest because of potential consequences, or because there are no consequences. Surveys are usually anonymous, but people still lie on them because they want to look good. The desirability bias causes people to lie about who they are relative to who they want to be. Impersonal data tends to be more honest. People will confess when they are alone rather than in the presence of others. Facilitates knowing what people do rather than say they do.
Big data facilitates better understanding of the topics which can lead to better resolution methods. Big data reveals that individuals are not alone in their insecurities and embarrassing behavior. Making overt covert suffering. Google data can highlight many vulnerable people, as they might not want to report their trauma to official sources.
Offline experiments are time consuming and costly. The digital space enables cheap and fast randomized experiments. Gaps in understanding can be filled by testing. Gaps always exist.
Big data has limitations. The numbers measure what can gathered, not necessarily what is wanted or important. Models created from the data does not indicate the reason why the model works. Knowing why models work may not be that important. But with this limitation, there is no indication of insights that can be gained and ways to improve understanding of the topics. There are data sources in which do facilitate lying rather than honesty. When there is no incentive to tell the truth, people make themselves appear better. Online presence is not always anonymous, and can cater to an audience.
Caveats?
The book is well written with plenty of examples and provides a general understanding of the power and complexity of big data. There are many topics in this book which are very sensitive, as in very private and personal. As such, the book may not be appropriate for minors.
The book is well written with plenty of examples and provides a general understanding of the power and complexity of big data. There are many topics in this book which are very sensitive, as in very private and personal. As such, the book may not be appropriate for minors.
Although big data does open up more opportunities to consider how people think, what matters is how the data is interpreted. There are a variety of interpretations of the data, of which there can be many misleading interpretations. Big data does offer lots to think about, but not how to think about what it brings up.
Content Warnings
Moderate: Sexual content
There are many topics in this book which are very sensitive, as in very private and personal. As such, the book may not be appropriate for minors.
I loved this book. It is definitely in my top 5 books of 2019, I cannot recommend it enough. I read some of the other goodread reviews and no one else seemed to have as good a time as me, so take this review with a grain of salt i guess.
Basically this book is giving you tangible data about how google probably knows you better than most of your friends and family. I don't really want to spoil much as I truly recommend the read! The data he provided was super interesting to me and proves how much you can gleam from internet searches. Topics covered range from voter turnout, election results, sexual preferences, methods of speech, sports and some other topics I'm sure I'm forgetting.
As always with data scientist, they don't seem to worry too much about the ethics of what they're doing, the author covered ethical questions in a mere 10 pages but I get that that's not the point he's trying to make so I'll give him a pass. The author is successfully explains how the social sciences will be turning more and more to data rather than surveys or small experiments when you can get access to pools and pools of data from Facebook, Google and Pornhub. He goes into what makes good research a bit (what to do with huge amounts of data etc) - if that's what you're after its quite sparse in terms of how to actually conduct good research but the results he shared were fascinating.
Also it's an easy read! Read it in 4 days, it's the holidays what do you have to lose!
PS. Appaz lots of people don't finish non fiction books?? By his standard fewer than 10% make it to reading the conclusion (he was looking at economics books but still) ?? Wild times man
A somewhat breezy, entertaining collection of stories and speculations regarding unexpected statistical correlations. It reminds me of and occasionally refers to Freakonomics. Stephens-Davidowitz's joy in his work comes across. I appreciate books like this that provide enough information on core concepts and methodology to give a sense of the rigors and significance of a practice but without a barrage of unexplained jargon and tedious details. (Those are what endnotes are for.) I also appreciate how data science explanations for non-specialists can highlight the impacts of innate and learned cognitive biases on rationality.