- Paperback: 302 pages
- Publisher: O'Reilly Media; 1 edition (August 14, 2015)
- Language: English
- ISBN-10: 1491916915
- ISBN-13: 978-1491916919
- Product Dimensions: 6 x 0.8 x 9 inches
- Shipping Weight: 12.6 ounces (View shipping rates and policies)
- Average Customer Review: 4.2 out of 5 stars See all reviews (9 customer reviews)
- Amazon Best Sellers Rank: #695,244 in Books (See Top 100 in Books)
Enter your mobile number or email address below and we'll send you a link to download the free Kindle App. Then you can start reading Kindle books on your smartphone, tablet, or computer - no Kindle device required.
To get the free app, enter your mobile phone number.
Creating a Data-Driven Organization: Practical Advice from the Trenches 1st Edition
Use the Amazon App to scan ISBNs and compare prices.
Frequently bought together
Customers who bought this item also bought
About the Author
Carl Anderson is the Director of Data Science at Warby Parker in New York overseeing data engineering, data science, supporting the broader analystics org, and creating a data-driven organization. He has had a broad-ranging career, mostly in scientific computing, covering areas such as healthcare modeling, data compression, robotics, and agent based modeling. He holds a Ph.D. in mathematical biology from the University of Sheffield, UK.
Browse award-winning titles. See more
If you are a seller for this product, would you like to suggest updates through seller support?
Top customer reviews
The book covers the basics of a data-driven organization from data collection, data quality, data analytics & metrics design to insights and leadership that drive actions. These form the basis for a data-driven culture in which the analytics value chain - the flow of data to final impact on business - is realized. The author thoroughly explains the analytics value chain through interviews and examples from data scientists and analytics leaders in various industries. It is filled with practical advice from the industries. You'll learn what works and what doesn't, and how to build a data culture that works for your organization.
The book is well written and organized into bite-size pieces that are easy to understand. If you want to develop a data-driven culture with the desire to turn data into actions quickly and efficiently, you will find this book helpful and worth your time.
Anderson talks a lot about an "analytics value chain" (data drives reports, to deeper drives and further analyses that get to a decision maker who changes strategies). Anderson emphasizes the culture of the organization is very important in being able to implement this type of work, and recognizes that the culture shift may take a while to occur. He also covers storytelling with data, a/b testing, an ideal data-driven culture set-up, and privacy, ethics and risk. There are good number of references, although the visualization, decision making, and a/b testing are a little light. Another topic that is missing from the book is results-based accountability, which can be helpful for groups drowning in metrics or data and can't choose (or don't know how to) because there are so many potential options.
It's pretty conversationally written, so it's a pleasant enough read, and is definitely worth the time for leaders (or people who are managing up) to think about ways to really drive forth data to action.
I found this book ... conventional. Perversely, I suppose, I found it too data-centric! Data does NOT drive the organization. Data - or better, information - informs it, and the difference goes a lot further than subtleties. The author takes things a bit far, IMHO, when it comes to suggesting C-level roles for data and analytics.
Data management issues are legion and getting more complex all the time. You cannot fix, solve, correct, or compensate them for them in the ways suggested by the author. Mind you, that does not obviate the need to take many of the steps covered in the book. After all, if you have a need for mastered data in a particular context, you have the need. Addressing data quality, especially narrowly focused on an issue, will always be necessary.
But none of the data management topics in the book, and that's what they are, will really address the information needs of a large, complex, modern corporate environment except at the margins. The goals, beyond simple monitoring mechanisms like dashboards and KPIs (the bane of my existence) are sometimes put in an analytics framework. They might be articulated as descriptive, prescriptive, and predictive, for example. Reporting strategies. Models. Etc. All of these are a house of cards built on assumptions of ownership, control, quality, security, and enforcement that are increasingly irrelevant. Want hardening of the company data arteries? Put modern "data management" in place. If it doesn't consume your company's entire budget or drive you to the bin, it will leave you holding, in a few years, a figuratively unoccupied leash wondering where the dog went as everyone, every business process, every vendor system, every legacy situation, every new technology, and every bit data freely available everywhere and anywhere undermines and obviates your strategy.
Here's the heart of it. State. We increasingly need ALL of our data in every conceivable state, in motion or at rest, even temporary and error-ridden, in infinitely possible combinations. Oh, and the rest of the world's data too, in the same way. No matter what your data management activity perspective, the problems just became insurmountable. The fact that most data forms don't carry sufficient state or sematic information complicates things further.
Doubless, I'm preaching to the choir here. But wait, you say, how about the more limited usages mentioned earlier? Doesn't that validate the need for data management structure? Well, yes and no. I'll stipulate the value of (data) control and definition within the context of a business operational process. But recognize that NON-operational data use, process-bound or not, is made more difficult by those controls and definitions. I'll note in passing that non-documented and unofficial processes exist (sometimes properly) WITHIN the boundaries of a larger business process ... and it is often the data in THIS context that is wanted by someone over yonder! (Just ask your nearest auditor.)
Finally, consider how any particular data is becoming increasingly irrelevant. (Again, all of this in the context of a very large organization or corporation.) An insufficiently appreciated phenomenon of truly vast data availability is that data compensates for data. What if you had, as a vendor who recently presented at my firm has, 30,000+ data points per person, tracked within thousands and thousands of tiny geographic boundaries? Don't have this or that data? This bit is old or of suspicious quality? Don't have access to yet something else? Guess what ... you can impute that data with astounding (and known) statistical accuracy. Um, your PII data policies aren't gone - you're going to be held to them, possibly to criminal violation levels, but it's now security theater and better viewed as a cost burden rather than a control.
By all means, have at the activities in "Creating a Data Driven Organization." It is well-written and clear (hence the 4 stars). You really have to to be able to operate. Just don't expect it to actually rise to the title's promise in the modern world.
Parting shot - expect "the organization" to fragment, too.