Be Data Literate: The Data Literacy Skills Everyone Needs to Succeed By JordanMorrow, Kogan Page, 2021, 215 pp, $15.99
本书面向管理者，系统介绍如何培养员工的数据素养，避免创建数据驱动文化时的常见陷阱，强调结合直觉与数据做决策，而非教授具体统计或软件操作。
The purpose of Morrow's Be Data Literate is to assist organizations that are attempting to create data-driven cultures, providing a holistic guide to improve workforce literacy. The author notes throughout the book that employees, regardless of job title or department, need to be data literate, using both the human element, that is, intuition, along with data collected and organized by software, to make informed decisions. Practitioners are the target audience of this book. It is important to clarify that this book does not show the reader how to perform statistical operations nor how to use software. Different software packages are mentioned but there are no instructions, per se. Be Data Literate is focused on avoiding pitfalls when attempting to create a data driven culture: its purpose is not to teach the reader analytics skills. The major strength of Be Data Literate is that it employs a systems approach to creating a data literate workforce. The major flaw of this book is the prose: it has countless redundant phrases, rhetorical questions, and it is not concisely written, which may frustrate its intended audience. In the opening chapter the data skills gap of today is adequately addressed. First, the author describes the data driven world we live in and dives into the Internet of Things (IoT), which means the “connectedness of everything.” He then provides an example of a runner whose watch, that collects data on every aspect of the run like elevation or pace, provides mountains of data to pour over. In this IoT kind of world, where technology is collecting unfathomable amounts of data, most people do not have the skills to understand, let alone make decisions, in this complex environment. Not only is the skills gap a problem for employers but trying to instill a data literate culture is a difficult challenge. Typically, new software is implemented, like Qlik or Tableau, yet employees do not even know how the software works, much less know how the data is collected and what the visualizations mean. Long story short: organizations fail to become data literate because they lack a coherent strategy, leading into the four levels of analytics. According to the author, there are foul levels regarding data literacy and most firms get stuck on the first: descriptive analytics. Employees spend too much time on making charts and graphs look visually appealing to describe trends, for instance, yet do not investigate the why behind the numbers. Similar to how General Motors struggled to implement lean practices like Toyota in the 1980s due to failing to systematically change their culture (Helper & Henderson, 2014), which is not an example from the book but is relatable, managers cannot simply add new technology to create a data driven enterprise. To move past descriptive statistics, and into the most important phase, diagnostic analytics, managers must work with employees to explore the why behind the data. For instance, realizing that sales have increased over the last quarter (descriptive) is not enough insight to make meaningful decisions. Connecting the revenue increase to a new incentive program (diagnostic), though, can lead to effective policy changes. Software that can be used in the diagnostic phase are Power BI, Excel, Qlik, and Tableau. The third level of analytics is predictive modeling. Understanding a trend and the why behind it is important, but forecasting the impact that today's decisions have on future events, or how changing x's predict y, is the purpose of predictive analytics. Software used in this realm are R, Python, SAS, Alteryx and more. Taking it to the final level, prescriptive analytics is when the technology makes suggestions as to what direction should be taken based on the modeling. Software used at this level are SAP and SAS. In this section of the text the author gives rudimentary examples of each level of analytics and ends with an important warning: employees in various roles and at all levels need to be trained in analytics or plans to change the culture will fail. The good news, as the text moves forward, is that data literacy has a simple definition: the ability to read, work with, analyze, and communicate with data. Morrow quickly points out that one does not need to be a data scientist to be data literate. Employees in various organizational functions, such as marketing, sales, and the executive team, can and should understand data. That does not mean they need to code in R, but it does mean they need competence in the four analytics levels. Many examples are provided regarding literacy, but they are all broad and none particularly novel. This section of the text is where the reader may become frustrated as the quantity of rhetorical questions starts to add up with no real payoff. For instance, each example begins in a similar fashion like: “first off, let's look at the information technology team. Do they need to work with data…of course they need to work with data…the information technology team works with data in many, many ways.” No specifics are provided, and this method of writing is prevalent throughout the text. Moreover, many sentences end and start exactly the same, that is, “…the executive group. The executive group.” Having noted the grammatical troubles of the text, there are nuggets of relevant information to be found. The author notes that people identify and respond to stories. If users can communicate their analyses within the framework of a story, they will be more effective. Although this relevant piece of information is present in the chapter it requires wading through too much hemming and hawing. Chapter four adds on to the data literacy discussion by describing the different tools and specifications under the data literacy umbrella: data and analytical strategy, data science, data ethics and regulation, data visualization, executive teams, data governance, culture, and data quality. The chapter describes each of these elements and makes two important points. First, departmental silos may emerge when implementing a data literate culture and a coherent strategy is needed to holistically create said culture. Second, he makes a case for STEM being STEAM, adding “art” to the acronym as data driven decisions should not purely be machine driven. Human creativity adds value to the process. Unfortunately, neither of the visual graphics in this chapter land. The first is the data literacy umbrella itself: it is just the previous terms listed under an umbrella graphic. Typically, visual models, like Maslow's Hierarchy of Needs, show how the terms relate to one another. Second, a heat map is presented to depict where a Cholera outbreak occurred, leading to the researchers identifying its root cause. Although this is an interesting discovery in the 1800s, it seems like a modern example could add value. The final visual is a line chart that shows the population growth of three different states. This is a decent example of how a trend can be visualized but the target audience will already be familiar with this type of chart. Regarding data literacy, the following chapter describes data fluency, the ability to speak and understand the language of data. Morrow recommends creating a data dictionary to ensure all employees are using the same terminology. Again, the text's main strength is that it makes a case for strategically, in this case standardizing the language, implementing a data literature culture. Yet, it stumbles with a visual that does not make complete sense. In this case, a graphic representing how data is communicated, in a circular flow, is not necessarily correct. Do data scientists only communicate with the workforce and analysts only, and not decision makers or the executive team? Interweaving of arrows would be more appropriate. At this point the text attempts to combine the concepts of data literacy and the four levels of analytics. A rehash of the levels is provided with new examples. However, data literacy terms are not combined with the four levels in a meaningful way. For instance, descriptive analytics is addressed but what are the common literacy issues here? Is it reading histograms and communicating their central tendency and dispersion? The text generically notes the literacy categories (i.e., communicating with data) at each of the four levels, which is a good starting point for managers, but not much more is included here. Now that leaders have a framework from which to work, the text transitions into the learning process. Morrow begins by stating that cultural change needs to happen with both a top down and bottom up/grassroots approach. Leadership can work with employees who will promote the change in an evangelistic way. Morrow's approach here is reminiscent of Daft's (2008) zoysia analogy where he suggests that employees supportive of cultural change should be strategically planted to sway those who are not, as is the case with how farmers plant zoysia plugs to overtake weeds. Employee skill levels should also be noted so skills gaps can be addressed. A few steps to cultural change are included by Morrow such as being transparent, creating a learning environment, and adopting a mentoring system. Regarding soft skills, the three C's of data literacy, curiosity, creativity, and critical thinking, are noted with examples provided, albeit many are outdated. A robust dashboard is presented regarding the curiosity component, which is a relevant visualization. The book begins its descent by noting that its intention is to make workforces more “data literate,” not “data driven.” Industry uses “data driven” terminology, at the cost of negating human experience and intuition. Data literacy attempts to use data to assist in decision making without managers blindly following software prescriptions. The author then presents a Data Informed Decision-Making Model, a feedback loop similar to the Plan, Do, Check, Act cycle, but it includes Ask, Acquire, Analyze, Integrate, and Decide. Once each part of the model is explained, with examples, the book ends with a quick overview of modern facets of analytics strategy: culture, business intelligence, artificial intelligence, machine learning and algorithms, big data, embedded analytics, the cloud, edge, and geo-analytics. Be Data Literate is a timely, practitioner-focused book that will help leaders avoid missteps while creating a data-competent culture. Although there are key takeaways from the text, such as breaking down silos and creating a data dictionary, the writing style will most likely alienate the reader.