What do you think?
Rate this book
Strong Ideas
Data Feminism
A new way of thinking about data science and data ethics that is informed by the ideas of intersectional feminism.
Today, data science is a form of power. It has been used to expose injustice, improve health outcomes, and topple governments. But it has also been used to discriminate, police, and surveil. This potential for good, on the one hand, and harm, on the other, makes it essential to ask: Data science by whom? Data science for whom? Data science with whose interests in mind? The narratives around big data and data science are overwhelmingly white, male, and techno-heroic. In Data Feminism, Catherine D'Ignazio and Lauren Klein present a new way of thinking about data science and data ethics—one that is informed by intersectional feminist thought.
Illustrating data feminism in action, D'Ignazio and Klein show how challenges to the male/female binary can help challenge other hierarchical (and empirically wrong) classification systems. They explain how, for example, an understanding of emotion can expand our ideas about effective data visualization, and how the concept of invisible labor can expose the significant human efforts required by our automated systems. And they show why the data never, ever “speak for themselves.”
Data Feminism offers strategies for data scientists seeking to learn how feminism can help them work toward justice, and for feminists who want to focus their efforts on the growing field of data science. But Data Feminism is about much more than gender. It is about power, about who has it and who doesn't, and about how those differentials of power can be challenged and changed.
Today, data science is a form of power. It has been used to expose injustice, improve health outcomes, and topple governments. But it has also been used to discriminate, police, and surveil. This potential for good, on the one hand, and harm, on the other, makes it essential to ask: Data science by whom? Data science for whom? Data science with whose interests in mind? The narratives around big data and data science are overwhelmingly white, male, and techno-heroic. In Data Feminism, Catherine D'Ignazio and Lauren Klein present a new way of thinking about data science and data ethics—one that is informed by intersectional feminist thought.
Illustrating data feminism in action, D'Ignazio and Klein show how challenges to the male/female binary can help challenge other hierarchical (and empirically wrong) classification systems. They explain how, for example, an understanding of emotion can expand our ideas about effective data visualization, and how the concept of invisible labor can expose the significant human efforts required by our automated systems. And they show why the data never, ever “speak for themselves.”
Data Feminism offers strategies for data scientists seeking to learn how feminism can help them work toward justice, and for feminists who want to focus their efforts on the growing field of data science. But Data Feminism is about much more than gender. It is about power, about who has it and who doesn't, and about how those differentials of power can be challenged and changed.
326 pages, Kindle Edition
First published March 17, 2020
383 people are currently reading
7072 people want to read
About the author
Catherine D'Ignazio
3 books8 followersCatherine D'Ignazio is Assistant Professor of Urban Science and Planning at MIT and coauthor of Data Feminism (MIT Press).
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 - 30 of 189 reviews
January 10, 2021
The authors of this book want us to be aware of our positionality, so I'll tell you that I've been a scientist in the field of genetics for over 20 years, more recently as a curator and manager of delivering data to the public. I'm also a woman of color who thinks a lot about ethics and equity. Which is to say, I really try to think critically about data, scientific and otherwise.
I take this book literally. I fully agree that we need to look at what biases lay behind the hypotheses and interpretations of studies. I think scientists need to be clear in what parameters and variables are actually being measured, and whether it's meaningful to use certain categories at all (for instance, in my field, US racial categories are not very meaningful, and their use often obscures conclusions from data). I think it's important to define your terms, be as precise as you can and aware of how shorthand labels may mislead. I think we should consider what blindspots we may have, especially when doing research that impacts people and societies. And that interactions on an equal standing with people with different perspectives (professional and personal backgrounds) can help point out these blindspots. I think we should interrogate why certain jobs in STEM projects (e.g. curator vs data scientist) have more or less prestige, and what that says about our society.
Therefore, I largely agree with the recommendations, such as they are, of this book.
But I can't take this book seriously, because it frames these concepts, which I consider simply good practice, as feminism (by feminism, they mean an intersection of a gender+racial+class justice lens). Maybe I'm too much of a consequentialist, but my kneejerk reaction to this book was to reject putting the fetishization of identity ahead of practice.
I was also hoping for more insights that would help in my professional life. Disappointingly, rather than give precise scenarios of study and analysis design, the authors tend to rely on quotes from feminist scholars they think are cool, and then shout out examples of research they like (i.e. done by people who they approve of) versus stuff they don't like (done by people they don't like).
I kept thinking of data scientists I work with, many of them older men who didn't grow up in the US, and how I would love to recommend a book like this to them, so that they would think deeper about responsible use of data. But this book doesn't provide an accessible entry into this subject.
I know the authors workshopped this manuscript - I just wish that in addition to consulting gender and cultural scholars, they had also been more intentional about running this work past more normies, maybe some of the ones who they mention were initially hostile to their project. Critical comments from that perspective would likely have resulted in a stronger product.
I take this book literally. I fully agree that we need to look at what biases lay behind the hypotheses and interpretations of studies. I think scientists need to be clear in what parameters and variables are actually being measured, and whether it's meaningful to use certain categories at all (for instance, in my field, US racial categories are not very meaningful, and their use often obscures conclusions from data). I think it's important to define your terms, be as precise as you can and aware of how shorthand labels may mislead. I think we should consider what blindspots we may have, especially when doing research that impacts people and societies. And that interactions on an equal standing with people with different perspectives (professional and personal backgrounds) can help point out these blindspots. I think we should interrogate why certain jobs in STEM projects (e.g. curator vs data scientist) have more or less prestige, and what that says about our society.
Therefore, I largely agree with the recommendations, such as they are, of this book.
But I can't take this book seriously, because it frames these concepts, which I consider simply good practice, as feminism (by feminism, they mean an intersection of a gender+racial+class justice lens). Maybe I'm too much of a consequentialist, but my kneejerk reaction to this book was to reject putting the fetishization of identity ahead of practice.
I was also hoping for more insights that would help in my professional life. Disappointingly, rather than give precise scenarios of study and analysis design, the authors tend to rely on quotes from feminist scholars they think are cool, and then shout out examples of research they like (i.e. done by people who they approve of) versus stuff they don't like (done by people they don't like).
I kept thinking of data scientists I work with, many of them older men who didn't grow up in the US, and how I would love to recommend a book like this to them, so that they would think deeper about responsible use of data. But this book doesn't provide an accessible entry into this subject.
I know the authors workshopped this manuscript - I just wish that in addition to consulting gender and cultural scholars, they had also been more intentional about running this work past more normies, maybe some of the ones who they mention were initially hostile to their project. Critical comments from that perspective would likely have resulted in a stronger product.
January 4, 2021
This book ignores the ideology that fuels data today. In a flawless feminist liberal approach, structural oppression is characterized like a "privilege hazard," ignoring a serious analysis of the economic framework where technology is situated today to fortress those oppressions.
November 5, 2020
Let me begin this review by paraphrasing something the authors express early on in the book: don’t let the title dissuade you; this is a book that everyone should pick up (and the authors facilitate this by making it freely available online). A true understanding of feminism and the values that define it has become distorted by contesting voices or is resisted by the uninitiated for various subjective reasons (often due to misunderstanding). With this in mind, this book proved to be a very helpful primer for key concepts within the discourse - intersectionality, privilege hazard, positionality and so on - and it also framed and explained scary sounding things like ‘the matrix of domination’. D’Ignazio and Klein demystify these and other ideas by providing their epistemic context and then grounding them using real world examples - and plenty of them. They discuss the seven principles of data feminism in great detail through the chapters. It is admirable how they lead by example: proceeding reflexively, applying each principle to themselves, and evaluating their positions in the social structure and relative to the subject matter. For them this text is only meant to serve as a starting point to further conversation and is not an end in itself.
Turning to its relevance, this book deftly deals with what is perhaps the most powerful item in the modern civilisational toolkit: data. What is it? Who collects it? For what purpose? With what consequences? Each chapter uses a particular lens to focus and understand what lies behind the curtain and explores the myriad complications involved with collecting and using data unquestioningly. At the same time, the point regarding the risks of the assumed objectivity of data is powerfully driven home. The authors highlight actionable steps to counter these and other problems at each stage.
Anyone working with data - whether its collection, design, ethics, visualisation or analysis - or with its potentially transformative application in the fields of public policy, development, activism, local community work etc, must consider this book an essential reading.
Lastly, to even a layperson who simply wants to better understand the way data shapes our world, to reflect on their own position within it, and to those who want to make it a little better, I cannot recommend this book enough. To my friend who perhaps thought I fit one of these categories, I’m grateful for it. (I also had her review this review to ensure that I’ve been responsible and accurate in my entry.)
Turning to its relevance, this book deftly deals with what is perhaps the most powerful item in the modern civilisational toolkit: data. What is it? Who collects it? For what purpose? With what consequences? Each chapter uses a particular lens to focus and understand what lies behind the curtain and explores the myriad complications involved with collecting and using data unquestioningly. At the same time, the point regarding the risks of the assumed objectivity of data is powerfully driven home. The authors highlight actionable steps to counter these and other problems at each stage.
Anyone working with data - whether its collection, design, ethics, visualisation or analysis - or with its potentially transformative application in the fields of public policy, development, activism, local community work etc, must consider this book an essential reading.
Lastly, to even a layperson who simply wants to better understand the way data shapes our world, to reflect on their own position within it, and to those who want to make it a little better, I cannot recommend this book enough. To my friend who perhaps thought I fit one of these categories, I’m grateful for it. (I also had her review this review to ensure that I’ve been responsible and accurate in my entry.)
June 3, 2020
4.5* stars but rounding to 5 because this is an essential read nowadays.
It was a pleasurable read. I learnt a lot along the way. And having synchronised the reading with the weekly video seminars with the authors made it a much pleasant experience than I had before with a more "academic" book.
Having experienced data science from my studies and as a job, I could relate myself a lot in this book. I learnt what is really behind data cleaning, why more women are hired as a data analyst (like myself) instead of data scientist, and which views are usually prioritised in data collection and analysis. But now, thanks to including the Feminism perspective, I see it as a field with high hopes for the future. As with many other areas, data science can be used for capitalistic purposes, but it can also help to bring up voices which have been silent for too long, visualise some deep problems and this way reaching a vast audience.
It was a pleasurable read. I learnt a lot along the way. And having synchronised the reading with the weekly video seminars with the authors made it a much pleasant experience than I had before with a more "academic" book.
Having experienced data science from my studies and as a job, I could relate myself a lot in this book. I learnt what is really behind data cleaning, why more women are hired as a data analyst (like myself) instead of data scientist, and which views are usually prioritised in data collection and analysis. But now, thanks to including the Feminism perspective, I see it as a field with high hopes for the future. As with many other areas, data science can be used for capitalistic purposes, but it can also help to bring up voices which have been silent for too long, visualise some deep problems and this way reaching a vast audience.
May 10, 2021
This is a great read, overall -- thoughtful conceptual framework, well-used theory, a great message and a lively tone. I certainly found some beautiful lines/quotations and great new formulations for how to think & speak about these topics in my workplace --
My main question mark about it is that I'm not entirely sure who the audience is. If it's people with a background in feminist/intersectionality theory, it's fundamentally pretty basic info (although the examples used to illustrate them are certainly interesting, albeit a few have that undergrad "I randomly picked something to use as an example in my paper and it's due tomorrow so we're just going with it ok" vibe). If it's data scientists who don't care about intersectionality, my suspicion is that it won't convince them. But if you're somewhere in the middle, or on one side with an interest in the other, I'm sure you'll pick at least something useful up!
My main question mark about it is that I'm not entirely sure who the audience is. If it's people with a background in feminist/intersectionality theory, it's fundamentally pretty basic info (although the examples used to illustrate them are certainly interesting, albeit a few have that undergrad "I randomly picked something to use as an example in my paper and it's due tomorrow so we're just going with it ok" vibe). If it's data scientists who don't care about intersectionality, my suspicion is that it won't convince them. But if you're somewhere in the middle, or on one side with an interest in the other, I'm sure you'll pick at least something useful up!
January 7, 2024
celui-là c'était vraiment pour l'école
j'ai même pas réalisé que je suis passé au travers en moins d'une semaine. c'était plaisant de le lire en buvant des cafés vanille-française d'la machine cheap du MIL pis de passer l'après-midi à coder et à "lire" des pages et des pages de codes incompréhensibles en buvant d'autres cafés vanille-française de la machine cheap du MIL. ça rendait la deuxième partie de la journée un peu plus digeste mettons. sans nécessairement en retirer quelque chose de théorique - ça se veut davantage pamphlétaire - ça reste que c'est un bel exemple des différentes manières d'approcher et d'écrire de manière située et féministe sur les données (numériques) pis la programmation. c'était aussi remplie de beaux exemples de data visualization.
j'ai même pas réalisé que je suis passé au travers en moins d'une semaine. c'était plaisant de le lire en buvant des cafés vanille-française d'la machine cheap du MIL pis de passer l'après-midi à coder et à "lire" des pages et des pages de codes incompréhensibles en buvant d'autres cafés vanille-française de la machine cheap du MIL. ça rendait la deuxième partie de la journée un peu plus digeste mettons. sans nécessairement en retirer quelque chose de théorique - ça se veut davantage pamphlétaire - ça reste que c'est un bel exemple des différentes manières d'approcher et d'écrire de manière située et féministe sur les données (numériques) pis la programmation. c'était aussi remplie de beaux exemples de data visualization.
November 25, 2020
A book called "Data Feminism" that suggests collecting data on sex can be and often is inappropriate...🥴. If you like to hem and haw and not land anywhere firm that might require you to say what you think this is absolutely the book for you.
September 20, 2021
After being assigned bits and pieces of this text in class, I finally bit the bullet and read it start to finish (mainly via the Open Access version from MIT Press).
I'll start by saying that a lot of critiques I'm seeing on Goodreads here are a bit unfounded. Many of them center around the idea that this book is wishy washy and lacks concrete steps or information that could be handed to data scientists. I think folks who are making this critique are unfamiliar with (or uninterested in) the ways in which feminism is highly dependent on context. That does not mean that this book lacks clear direct guidance. (I mean, each of the chapters is literally organized a principle for action, backed up by feminist theory with examples of that principle in data science projects?). It means that you're going to have to take a general principle, like challenge power, and apply it to your specific context. That's going to take work and research and energy. If you're expecting this book to give you easy steps that you can check off (and then treat yourself to an ally cookie), look elsewhere.
I agree that perhaps Data Feminism could have dug a bit deeper into the ways that capitalism and data aggregation are intertwined. But to say that the book ignores these connections is also a bit far fetched, especially considering they have a whole chapter dedicated to the concept of invisible labor and how it connects to the data and tech industries. Folks who were craving more (and deeper connections) between data and capitalism ought to check out Platform Capitalism or any one of the many, many books mentioned throughout this book.
I personally have found this book to be an excellent starting point for how to think about, share, and connect communities to data (especially when thinking about my role as a data intermediary). The examples are numerous, rich, and detailed (I just wish the open access version had hyperlinks embedded in images to link to those projects!).
I'll start by saying that a lot of critiques I'm seeing on Goodreads here are a bit unfounded. Many of them center around the idea that this book is wishy washy and lacks concrete steps or information that could be handed to data scientists. I think folks who are making this critique are unfamiliar with (or uninterested in) the ways in which feminism is highly dependent on context. That does not mean that this book lacks clear direct guidance. (I mean, each of the chapters is literally organized a principle for action, backed up by feminist theory with examples of that principle in data science projects?). It means that you're going to have to take a general principle, like challenge power, and apply it to your specific context. That's going to take work and research and energy. If you're expecting this book to give you easy steps that you can check off (and then treat yourself to an ally cookie), look elsewhere.
I agree that perhaps Data Feminism could have dug a bit deeper into the ways that capitalism and data aggregation are intertwined. But to say that the book ignores these connections is also a bit far fetched, especially considering they have a whole chapter dedicated to the concept of invisible labor and how it connects to the data and tech industries. Folks who were craving more (and deeper connections) between data and capitalism ought to check out Platform Capitalism or any one of the many, many books mentioned throughout this book.
I personally have found this book to be an excellent starting point for how to think about, share, and connect communities to data (especially when thinking about my role as a data intermediary). The examples are numerous, rich, and detailed (I just wish the open access version had hyperlinks embedded in images to link to those projects!).
March 22, 2021
Ironically, this book is potentially more harmful than something which is less correct in its diagnosis of the problems associated to data science and data collection writ large.
This book repeatedly gets lots of things right, in terms of describing the problems that we see: the biases associated to data collection, the biases inherent in having machine learning systems developed by unrepresentative groups, the ways in which systems and data collection can reinforce existing (malicious) power structures, the racist / sexist epistemic assumptions baked into data collection / analysis, etc.
It even goes so far as to make correct claims about the politics: the ineffective nature of white saviorism, education as not necessarily an emancipatory force, the existence of power dynamics that preclude political action, etc.
Then what?! It ultimately leaves any of the broader structural analysis at the roadside (racial capitalism, the particular history and economics of the way that data science is used in our modern capitalist system, etc.), and gives a broad guide about how "we" ought to be teaching data science. It falls back to a depressingly-liberal theory of change after hitting the nail on the head about all sorts of stuff, even going so far as to quote people that _certainly_ (and correctly) had more radical prescriptions about what to do with all this.
The word "capitalism" appears exactly 3 times in this entire 200 page text, and only in passing, which might just summarize the problem with the whole damned thing. If it were wrong, we could maybe dismiss it more easily, and therein lies the danger.
This book repeatedly gets lots of things right, in terms of describing the problems that we see: the biases associated to data collection, the biases inherent in having machine learning systems developed by unrepresentative groups, the ways in which systems and data collection can reinforce existing (malicious) power structures, the racist / sexist epistemic assumptions baked into data collection / analysis, etc.
It even goes so far as to make correct claims about the politics: the ineffective nature of white saviorism, education as not necessarily an emancipatory force, the existence of power dynamics that preclude political action, etc.
Then what?! It ultimately leaves any of the broader structural analysis at the roadside (racial capitalism, the particular history and economics of the way that data science is used in our modern capitalist system, etc.), and gives a broad guide about how "we" ought to be teaching data science. It falls back to a depressingly-liberal theory of change after hitting the nail on the head about all sorts of stuff, even going so far as to quote people that _certainly_ (and correctly) had more radical prescriptions about what to do with all this.
The word "capitalism" appears exactly 3 times in this entire 200 page text, and only in passing, which might just summarize the problem with the whole damned thing. If it were wrong, we could maybe dismiss it more easily, and therein lies the danger.
October 18, 2020
This is a fantastic introductory text for several fields— most notably intersectional feminism, data ethics/justice, and visual cultures. D'Ignazio and Klein dive into the highly relevant dangers that unmonitored data can cause, and use concrete and timely examples to ground their theories in reality. Instead of pointing out a problem and despairing, they delineate actionable steps relevant to anyone using digital technologies, even for the average "layperson" who just use sites such as Facebook or Twitter. This is a crucial text to anyone interested in what goes into data creation, maintenance, and/or use. This is a crucial text for anyone curious about how data manipulation impacts your day to day life. This is a crucial text for anyone interested in data activism.
D'Ignazio and Klein are in the privileged position to permit the free use of their book, and I recommend taking advantage of that! Their book is free to read online. I would especially recommend reading their Intro, Chapter 3, and Chapter 6 at the bare minimum.
D'Ignazio and Klein are in the privileged position to permit the free use of their book, and I recommend taking advantage of that! Their book is free to read online. I would especially recommend reading their Intro, Chapter 3, and Chapter 6 at the bare minimum.
May 5, 2023
Don't disagree with the anti- capitalist agenda of this set of essays...also sympathetic to the feminist ideology and connecting it to tech work. However, the essay collection as a whole feels one-dimensional and "preaching to the choir"-esque. Standard rhetorical strategy gives at least some space and effort to elucidate the opposing side of these viewpoints and the writers take for granted that doing so would strengthen all the supposedly supportive data they infuse throughout the chapters. Maybe it's just the fact that these viewpoints have, over the past three years, become mainstream...? There's a so-what? quality to the overarching points that data requires context and workers of color are not getting their fair share (which is also kind of supporting capitalism...? Like, the only way to succeed is to give money to ppl, which is true, but then fails to move society beyond capitalism...tricky).
Idk...I feel like I'm on board but anyone who doesn't already believe in these ideas already wouldn't be convinced, so it's a bit disappointing in that sense.
Idk...I feel like I'm on board but anyone who doesn't already believe in these ideas already wouldn't be convinced, so it's a bit disappointing in that sense.
December 14, 2020
I used this book as a textbook in my undergraduate Introduction to Digital Curation class. Based on the discussion and written assignments I think the students found it engaging and informative. The format, illustrations and overall structure of the book make for a very compelling read. It also includes very generous citations to the research literature and relevant projects which serve as a gateway into a wider world of cirtical data studies. I think this means it would work well in a digital visualization or digital humanities class as well.
April 9, 2021
4.5 Stars.
This is an inspirational book. So glad that this book have introduced me to the 7 data feminism principles, as well as all the wonderful projects around the world, especially those community-driven projects! They gave me lots of ideas.
And I love how the authors audit their citations, making their greatest efforts to remain reflexive.
This is an inspirational book. So glad that this book have introduced me to the 7 data feminism principles, as well as all the wonderful projects around the world, especially those community-driven projects! They gave me lots of ideas.
And I love how the authors audit their citations, making their greatest efforts to remain reflexive.
July 7, 2021
I gained an education on how data is used and abused in service of power, and how to think clearly enough about it to do things right and better. Really enjoyed this book and will use its lessons to design more ethical, inclusive, and just systems, products, processes, and companies.
March 20, 2024
3.5 stars from me!
Really interesting perspective relating to data visualisations, and how keeping things "neutral" sends a message itself.
I loved all the content relating to how the whole data process shouldn't be limited to "experts" and that we should engage with the communities we are collecting data on and that they should be involved. Working in government we consistently collect data that underrepresents certain groups and I've always thought this has a lot do with our eurocentric approach to data collection.
Didn't really enjoy the way it was written, but gave me a lot to think about in terms of the way I have been taught to analyse and decipher data.
Really interesting perspective relating to data visualisations, and how keeping things "neutral" sends a message itself.
I loved all the content relating to how the whole data process shouldn't be limited to "experts" and that we should engage with the communities we are collecting data on and that they should be involved. Working in government we consistently collect data that underrepresents certain groups and I've always thought this has a lot do with our eurocentric approach to data collection.
Didn't really enjoy the way it was written, but gave me a lot to think about in terms of the way I have been taught to analyse and decipher data.
March 28, 2022
Catherine D'Ignazio and Lauren Klein present an evolutionary point of view about data science and data ethics, informed by intersectional feminist thought.
With a structured and full of examples approach, the book offers strategies for data scientists seeking to learn how feminism can help them work toward justice, and for feminists who want to focus their efforts on the growing field of data science.
The ideas presented on this book can be vastly explored in many other challenges related to intersectional data ethics.
With a structured and full of examples approach, the book offers strategies for data scientists seeking to learn how feminism can help them work toward justice, and for feminists who want to focus their efforts on the growing field of data science.
The ideas presented on this book can be vastly explored in many other challenges related to intersectional data ethics.
December 31, 2021
5 stars
Phenomenal. Everyone should read this, irrespective of whether you work in data science. This is important, accessible, well and intentionally researched, and most importantly, extremely sensitively (but boldly) written. I loved it and I know for sure that it has changed the way I will approach and work with data from now on.
Phenomenal. Everyone should read this, irrespective of whether you work in data science. This is important, accessible, well and intentionally researched, and most importantly, extremely sensitively (but boldly) written. I loved it and I know for sure that it has changed the way I will approach and work with data from now on.
May 3, 2021
A book full of convincing and interesting data projects and examples of how the many elements of data can be manipulated to serve the users purpose. The authors take an intersectional feminist lens and applies it to the tech industry. Its arguments are solid, but I disagreed with many of their suggestions. So much of the academic literature has shown increasing diversity among teams doesn't increase performance and replacing the objectivity of science with subjective experience doesn't sit right with me.
March 16, 2021
This is a thorough and insightful exploration of the ways big data can perpetuate or exacerbate power imbalances when the people collecting, analyzing and reporting the data fail to interrogate the processes behind their measurement. It's well worth a read for anyone who uses or plans to use data as a tool for examining social structures – i.e., anyone in the sciences, the humanities or journalism – or anyone who might be reading those kinds of analyses. Which I guess is most everyone.
March 20, 2022
have you ever thought whether your dataset is *really* objective? or whether it perpetuates (not so) subtle discrimination and preserves the unequal status quo in its use and how the data are being displayed or represented? are data racist?
---
those questions might sound, uhm, silly. but it’s actually not—and they’re legitimate concern. have you ever thought how is the process of data collection? is data really n dimensional—and it most definitely excludes societal and cultural context? what about the labor costs of cleaning them, and invisible hands which went uncredited?
data feminism is actually pretty straightforward and literal. but how does it work? how is it possible to apply feminist principles when we’re working with data? D’Ignazio and Klein propose the seven core principles of data feminism: it begins with examining how power operates in the world today—who benefits from it, who are neglected and who are harmed?
“for people in positions of power and privilege, issues of race and gender and class and ability—to name only a few— are OPP: other people’s problems. In other words, those who occupy positions of privilege in society are able to remain innocent of that privilege. race becomes something that only people of color have. gender becomes something that only women and non-binary people have. sexual orientation becomes something that all people except heterosexual people have. and so on.”
by examining the power, we can quantify and visualize the structural oppression to challenge and change the existing power and status quo preserved in the world today.
third, is to elevate emotion and embodiment—honoring context while not subordinating ethocs nor emotions while visualizing them.
fourth, is to rethink binaries and hierarchies—reexamining and rethinking assumptions and beliefs behind the infrastructure classifications. are we doing more harm than good?
“in a world in which quantification always leads to accurate representation, and accurate representation always leads to positive change, then always counting gender identities outside the binary makes perfect sense. But being represented also means being made visible, and being made visible to the matrix of domination—which police the gender binary—poses significant risks to the health and safety of minoritized groups.”
fifth, is to embrace pluralism. data scientists are not heroes and the data are not *mere* data, they are lived experience, trying to be quantified and translated into reports—rather than regarding themselves as such, data for co liberation model is offered as a better replacement for data for good model. There is a transfer knowledge between the data/theory experts to communities, which cultivates solidarity between both parties.
We also need to consider context of how the data is being represented, which includes social power analysis of the said data setting. Lastly, we need to make labor visible—honoring the invisible labor involved in every step of data science. Whose labor is counted, and whose is being screened out?
Something that really sticks with me is this paragraph from the nth chapter:
“Big Dick Data is a formal, academic term that we, the authors, have coined to denote big data projects that are characterized by masculinist, totalizing fantasies of world domination as enacted through data capture and analysis. Big Dick Data projects ignore context, fetishize size, and inflate their technical and scientific capabilities.”
which reminds me of a famous indonesian politician, well-known for his political activism, but also for his bIg dAtA glorification 😛
I really appreciate their acknowledgement of their positions and experiences—of being white womxn from the global north, exposed to the elitist production of knowledge. it feels like multiple disclaimers that *they* are in the position of privilege, but also not an excuse to oversee anything. as the book employs the term “intersectional feminism”, they try to not talk exclusively about womxn and gender like other libfem books, but they try to incorporate positionality in every aspect of data science.
another thing that i find to be lovely is the accessibility. although sold exclusively as a hardback, you can also read the book online on their site, data-feminism.mitpress.mit.edu/. they also add a brief summary by the end of each chapter—which i find to be wonderful.
a few considerations into reading this: i still can feel that trans folks here are somewhat still being excluded, although they are being included in the metrics. in this book it is often written as “women and nonbinary people”—is it because they are trying to avoid the usage of “trans womxn/men” as they can be regarded as exclusionary as well?
two, is that the print format—which is only available in hardback—makes the book really uncomfortable to be read casually. i personally chose to lend the digital copy in my library, although i own a copy. with its square shape and fancy paper, this book is incredibly heavy for casual reading, the font is too small to be read casually before bed. if anyone wants to read this, i’d strongly suggest to grab the digital copy—be it the ebook or the online copy.
also, upon finishing the book i stumbled upon elizabeth garbee’s review that she was confused for whom was this book written—which i can agree that this lies in a confusing intersection: too “simple” for academics, too heavy for casual bookstore stroller. but like i said above, it’s only introductory level. in the introduction is also written that this book is meant for feminists trying to expand their knowledge within context of tech and data scientists to learn how feminism can be applied into their field by providing cases in which it was used—which can be understood why this book doesn’t seem so thorough.
since i want to avoid giving ratings—we all know that they’re subjective, but i want to give more context that **they’re good**—here’s my first attempt of rationalizing my judgement:
rounding up from my observations, i’d still recommend people to read this, especially those who are not quite familiar in the industry, or those are in the industry but quite unfamiliar with feminism. i haven’t finished invisible women nor algorithms of oppression, but from what i’ve read, if you enjoy both books, most likely you’ll enjoy this as well.
as for someone who only knows one or two things in data science, this book is definitely for me.
---
those questions might sound, uhm, silly. but it’s actually not—and they’re legitimate concern. have you ever thought how is the process of data collection? is data really n dimensional—and it most definitely excludes societal and cultural context? what about the labor costs of cleaning them, and invisible hands which went uncredited?
data feminism is actually pretty straightforward and literal. but how does it work? how is it possible to apply feminist principles when we’re working with data? D’Ignazio and Klein propose the seven core principles of data feminism: it begins with examining how power operates in the world today—who benefits from it, who are neglected and who are harmed?
“for people in positions of power and privilege, issues of race and gender and class and ability—to name only a few— are OPP: other people’s problems. In other words, those who occupy positions of privilege in society are able to remain innocent of that privilege. race becomes something that only people of color have. gender becomes something that only women and non-binary people have. sexual orientation becomes something that all people except heterosexual people have. and so on.”
by examining the power, we can quantify and visualize the structural oppression to challenge and change the existing power and status quo preserved in the world today.
third, is to elevate emotion and embodiment—honoring context while not subordinating ethocs nor emotions while visualizing them.
fourth, is to rethink binaries and hierarchies—reexamining and rethinking assumptions and beliefs behind the infrastructure classifications. are we doing more harm than good?
“in a world in which quantification always leads to accurate representation, and accurate representation always leads to positive change, then always counting gender identities outside the binary makes perfect sense. But being represented also means being made visible, and being made visible to the matrix of domination—which police the gender binary—poses significant risks to the health and safety of minoritized groups.”
fifth, is to embrace pluralism. data scientists are not heroes and the data are not *mere* data, they are lived experience, trying to be quantified and translated into reports—rather than regarding themselves as such, data for co liberation model is offered as a better replacement for data for good model. There is a transfer knowledge between the data/theory experts to communities, which cultivates solidarity between both parties.
We also need to consider context of how the data is being represented, which includes social power analysis of the said data setting. Lastly, we need to make labor visible—honoring the invisible labor involved in every step of data science. Whose labor is counted, and whose is being screened out?
Something that really sticks with me is this paragraph from the nth chapter:
“Big Dick Data is a formal, academic term that we, the authors, have coined to denote big data projects that are characterized by masculinist, totalizing fantasies of world domination as enacted through data capture and analysis. Big Dick Data projects ignore context, fetishize size, and inflate their technical and scientific capabilities.”
which reminds me of a famous indonesian politician, well-known for his political activism, but also for his bIg dAtA glorification 😛
I really appreciate their acknowledgement of their positions and experiences—of being white womxn from the global north, exposed to the elitist production of knowledge. it feels like multiple disclaimers that *they* are in the position of privilege, but also not an excuse to oversee anything. as the book employs the term “intersectional feminism”, they try to not talk exclusively about womxn and gender like other libfem books, but they try to incorporate positionality in every aspect of data science.
another thing that i find to be lovely is the accessibility. although sold exclusively as a hardback, you can also read the book online on their site, data-feminism.mitpress.mit.edu/. they also add a brief summary by the end of each chapter—which i find to be wonderful.
a few considerations into reading this: i still can feel that trans folks here are somewhat still being excluded, although they are being included in the metrics. in this book it is often written as “women and nonbinary people”—is it because they are trying to avoid the usage of “trans womxn/men” as they can be regarded as exclusionary as well?
two, is that the print format—which is only available in hardback—makes the book really uncomfortable to be read casually. i personally chose to lend the digital copy in my library, although i own a copy. with its square shape and fancy paper, this book is incredibly heavy for casual reading, the font is too small to be read casually before bed. if anyone wants to read this, i’d strongly suggest to grab the digital copy—be it the ebook or the online copy.
also, upon finishing the book i stumbled upon elizabeth garbee’s review that she was confused for whom was this book written—which i can agree that this lies in a confusing intersection: too “simple” for academics, too heavy for casual bookstore stroller. but like i said above, it’s only introductory level. in the introduction is also written that this book is meant for feminists trying to expand their knowledge within context of tech and data scientists to learn how feminism can be applied into their field by providing cases in which it was used—which can be understood why this book doesn’t seem so thorough.
since i want to avoid giving ratings—we all know that they’re subjective, but i want to give more context that **they’re good**—here’s my first attempt of rationalizing my judgement:
rounding up from my observations, i’d still recommend people to read this, especially those who are not quite familiar in the industry, or those are in the industry but quite unfamiliar with feminism. i haven’t finished invisible women nor algorithms of oppression, but from what i’ve read, if you enjoy both books, most likely you’ll enjoy this as well.
as for someone who only knows one or two things in data science, this book is definitely for me.
January 16, 2022
This book should be a required text for every data scientist and quantitative social scientist. While I think there are some things that could have been more theoretically well explained, in the context of what this book is attempting to do, it wasn't necessary to making its point.
This book is possibly the most valuable, usable, and accessible read I've seen explaining the ground rules for using data to overcome oppression and to grow knowledge which "co-liberates" researchers, those stuck in systematically oppressive tasks, and those harmed by oppressive actions.
I do have a minor rhetorical critique with the way this book discusses and qualifies oppression. The book could be read in a way which uncritically trusts all emotional reactions to data as a kind of oppression. While I think there is some labor by the authors to show how to avoid this, on page 57, the inductive narrative is that people know they are being oppressed before the data shows it, which is true.
There is a rise in critical scholarship showing how there are false positives with this kind of induction that should be taken seriously. For example, often Evangelical Christians see themselves as an oppressed group when they actively are enraged by not having the right to be the policy makers, the moral gatekeepers, i.e. the oppressors. Also there are many examples of people "feeling" other people's pain in research without material connection to it, and thus the actual pains happening are misaccounted, misdiagnosed, and cause even deeper harms. Similarly, there's a recent trend of discussing what's been called the oppression of "white woman tears" which also should be examined closely in these contexts.
While the authors seem to have no intention of empowering potential false positives, one could easily use this book's actionable methodology of doing data science to so this. So it would have been nice to see a little more about how to avoid these cases in more appropriate locations.
Additionally, while I understand the intention of the sections on data visualization, some of the discussion there pushes towards arguing towards possibly unethical uses of data manipulation. The book explicitly justifies how data visualization and data sets are not neutral and attempts to make them appear "objective" are merely an aesthetic of "objectivity" which is a psychological trick rather than actual objectivity. I agree with all of this, but I think more could have been shown to qualify the differences between the use of emotion in data visualization for scientific good and data manipulation. Is there a good place where we can draw the line? Where is the line between appealing to emotions and psychological manipulation? The book appears to want us to read it in good faith here and hopes that people will know where the moral line is. Perhaps we don't actually know yet, but this book opens these questions for me and I wished there were answers or at least a caveat about misuse of data visualizations.
This book is possibly the most valuable, usable, and accessible read I've seen explaining the ground rules for using data to overcome oppression and to grow knowledge which "co-liberates" researchers, those stuck in systematically oppressive tasks, and those harmed by oppressive actions.
I do have a minor rhetorical critique with the way this book discusses and qualifies oppression. The book could be read in a way which uncritically trusts all emotional reactions to data as a kind of oppression. While I think there is some labor by the authors to show how to avoid this, on page 57, the inductive narrative is that people know they are being oppressed before the data shows it, which is true.
There is a rise in critical scholarship showing how there are false positives with this kind of induction that should be taken seriously. For example, often Evangelical Christians see themselves as an oppressed group when they actively are enraged by not having the right to be the policy makers, the moral gatekeepers, i.e. the oppressors. Also there are many examples of people "feeling" other people's pain in research without material connection to it, and thus the actual pains happening are misaccounted, misdiagnosed, and cause even deeper harms. Similarly, there's a recent trend of discussing what's been called the oppression of "white woman tears" which also should be examined closely in these contexts.
While the authors seem to have no intention of empowering potential false positives, one could easily use this book's actionable methodology of doing data science to so this. So it would have been nice to see a little more about how to avoid these cases in more appropriate locations.
Additionally, while I understand the intention of the sections on data visualization, some of the discussion there pushes towards arguing towards possibly unethical uses of data manipulation. The book explicitly justifies how data visualization and data sets are not neutral and attempts to make them appear "objective" are merely an aesthetic of "objectivity" which is a psychological trick rather than actual objectivity. I agree with all of this, but I think more could have been shown to qualify the differences between the use of emotion in data visualization for scientific good and data manipulation. Is there a good place where we can draw the line? Where is the line between appealing to emotions and psychological manipulation? The book appears to want us to read it in good faith here and hopes that people will know where the moral line is. Perhaps we don't actually know yet, but this book opens these questions for me and I wished there were answers or at least a caveat about misuse of data visualizations.
Read
August 30, 2020I have published a long text on data feminism in my blog post which you can find here.
Also, you can find a much shorter version here. Just expect some spoilers (if you can spoil non-fiction) :)
Also, you can find a much shorter version here. Just expect some spoilers (if you can spoil non-fiction) :)
June 26, 2023
pleasantly surprised by this. there is a lot of intellectual handholding re: intersectional feminist values / methods BUT the introduction and description of ideas is effective and only at times reductive. the chapters are well constructed and a very readable primer for critical information studies for presumably a STEM audience. not sure about the citation audit at the end — I haven’t seen anything like it in other academic works I’ve read but its methods clash unpleasantly with its (self admitted) argument about the challenges of ANY identity-based classification system / audit. the authors do a fantastic history of the idea of “data” itself being rooted in ideology but it doesn’t quite escape the cycle by prioritizing this number-based citation audit instead of other formats.
if everything is data, then nothing is data, and is a data scientist the same thing as a historian by another name, imbued with different cultural meaning and value?
still, mostly fantastic. recommending to my mother ✅
if everything is data, then nothing is data, and is a data scientist the same thing as a historian by another name, imbued with different cultural meaning and value?
still, mostly fantastic. recommending to my mother ✅
July 16, 2025
I first came across this book in my cartography classes, but having gone through the whole thing, it feels fitting for anyone dabbling in the quantitative social sciences. They don’t rely on the cop out of only using qualitative methods to deliver ethical research, instead espousing 7 principles to approaching quantitative research that forwards feminist theory rather than hiding from it. The quant methods and feminist theory questions are so often separate when they do not need to be. While I think there are some fair criticisms of the feasibility of consistently utilizing all the principles in this book for every project in a world of neoliberal deadlines/quotas, I appreciate the spirit of their project and their message of how we can always try harder to keep that spirit in our work.
August 19, 2021
An insightful and eye-opening book on how we should use data for justice and to challenge the current power hierarchy and privilege operating in data science. The definition of feminism in the book title is beyond gender and stresses the importance of addressing intersectionality. I am especially impressed by the efforts of the authors and researchers in creating metrics to guide the writing process as an attempt to address structural problems such as racism and patriarchy by, for example, including examples and citations from people of colours and women. This is so much needed!
March 18, 2022
I loved so many of the projects featured in this book and felt inspired reading about them. I also particularly loved the book’s accountability metrics and audit notes as a model for how to use both citation and editing stages to further embed values into the work.
Overall, this book presents a great framework for how to use data in the new world:
Examine power
Challenge power
Elevate emotion and embodiment
Rethink binaries and hierarchies
Embrace pluralism
Consider context
Make labor visible
Boom.
Overall, this book presents a great framework for how to use data in the new world:
Examine power
Challenge power
Elevate emotion and embodiment
Rethink binaries and hierarchies
Embrace pluralism
Consider context
Make labor visible
Boom.
September 18, 2023
I read a lot, but I don’t often have the feeling right away that a book will drastically change the way I think about things. From the first chapters of Data Feminism I already noticed that I started thinking about data differently and that I wanted to share my insights with people around me. While strongly focused on the North American context, I immediately saw ways that I could apply the principles of data feminism in my own work and it more or less pulled me out of an academic slump that I didn’t really know I was in. Anyone working with data, please read this book. Please. 🙏🏻
January 18, 2024
3/5
I think this book's overall message is important but I didn't like how they separated the sections, they were all too similar. It's also pretty dense and reads a bit like a textbook at times. The two authors also don't add much value to this topic in my opinion, the majority of the book is them citing books, studies and events. I do think some of these studies and events were fascinating to read about.
I think this book's overall message is important but I didn't like how they separated the sections, they were all too similar. It's also pretty dense and reads a bit like a textbook at times. The two authors also don't add much value to this topic in my opinion, the majority of the book is them citing books, studies and events. I do think some of these studies and events were fascinating to read about.
March 20, 2024
audiobooked this so i feel like i barely internalized anything
in general all i can remember is
- intersectionality is super important as always..
- context for datasets and why they are missing, how they came about = super important
- the new jim code......... so bad
- data visualization design: how the story affects you and the emotion it can draw is important
to re-read more in depth once i have the bandwidth...
in general all i can remember is
- intersectionality is super important as always..
- context for datasets and why they are missing, how they came about = super important
- the new jim code......... so bad
- data visualization design: how the story affects you and the emotion it can draw is important
to re-read more in depth once i have the bandwidth...
September 25, 2020
i think everyone (especially those in research) should read this book. it's a great overview on the situated nature of data and what we can do to fight back against and avoid perpetuating injustices in data work. it's, of course, not comprehensive of all the ways systemic oppression shows up in data science, but I think this would pair nicely with further reading on decolonizing science/indigenous science.
Displaying 1 - 30 of 189 reviews