This chapter provides an overview of body image disorders as they pertain to men. Body image encapsulates thoughts, beliefs, and feelings about one’s physical appearance. For some men, these thoughts, beliefs, and feelings are neutral, or even positive. This is ideal, insofar as one’s body ought to be a functional and useful asset that allows an individual to live life on their own terms. Yet for others, these thoughts, beliefs, and feelings are decidedly negative.

Pharmacokinetics refers to the effects the body has upon a consumed drug, by considering a variety of processes which a dug undergoes during its time within the body. In contrast pharmacodynamics can be considered as the effects a drug has upon the body that has consumed it, by considering the drug’s effects at its principal sites of action. Safe and effective therapeutic management of drugs for individual patients requires application of pharmacokinetic and pharmacodynamic principles to enhance efficacy and minimize toxicity. Chapter 5 (Pharmacogenomics and Psychopharmacology) provides a detailed review of the genes relevant for drug absorption, distribution, metabolism and excretion. As such these issues are only dealt with briefly in this chapter.

This final chapter considers the future growth of data science and the role of information professionals as they continue to reposition themselves in the changing information ecosystem. Short of a worldwide catastrophe, it is impossible to see anything but the continued rapid growth of data over the coming years. Whatever measure is used, data production and consumption is expected to rise. For example, the Computer Information System Company (Cisco) predicts that by 2022 there will be more internet traffic per year than there was between 1984 and 2016 (Cooney, 2018).

The way particular sets of data are able to be used will differ. With the introduction of the GDPR access to personal data is already restricted, but whether we are talking big data, little data or some as yet unheralded type of data – there will be more data and more data science.

This growth will be accompanied by many opportunities for data science in the sciences, arts and humanities, as well as in homes, offices, businesses and of course libraries. Data science work has been described as the ‘sexiest job of the 21st century’ (Davenport and Patil, 2012), and while that is a description that is rarely levelled at library and information professionals, this book should have helped to persuade readers that there are opportunities for library and information professionals to help others apply data science methods, and to use data science methodologies in their own work.

This chapter consists of three parts:

• a list of eight challenges facing data science; although it is a field in the ascendency, associated with progress and all things modern, it is essential that we are aware of some of the challenges data science continues to face, and don't get caught up in the more extreme hyperbole of the hype cycle

• a list of ten steps that would-be library and information science professionals might take when embarking on a career in data science librarianship; as has been frequently expressed throughout this book, a lot falls under the data science umbrella, and the practical first steps are all too easily overlooked

• a short call to play; there is no book or course that can give more than the most basic of introductions to data science.

Today there are vast quantities and varieties of data, with new data sources emerging all the time. This makes any attempt to provide an overview of data a futile and quickly outdated task. Therefore this chapter considers some of the data that is available according to three features: whether it is ‘big data’, its format, and its source.

‘Big data’ is typically used to describe the large quantities of data now being generated and the complex infrastructure needed for their collection and analysis, and it can be contrasted with small data sets that can be simply gathered and analysed on a desktop. Data format provides an overview of some of the ways data can appear, from tables of data in documents through to APIs and linked data. Finally data sources looks more closely at some of the sources of data that are currently publicly available, as well as some of the additional data a library or information professional may have access to.

Big data

The term ‘big data’ can be traced back to the mid-1990s (Kitchin and McArdle, 2016), although it really entered the public consciousness in 2012. The New York Times ran articles with titles such as ‘The Age of Big Data’ (Lohr, 2012a) and ‘How Big Data Became So Big’ (Lohr, 2012b); the Guardian had ‘Why Big Data is Now Such a Big Deal’ (Naughton, 2012b) and ‘Big Data: revolution by numbers’ (Naughton, 2012a); and Big Data, Big Impact (WEF, 2012) was a topic at Davos in 2012. Google Trends shows there was a rapid rise in online searches for the term over the year, and since then there has been a raft of popular science and business publications devoted to the subject: Big Data (Mayer-Schönberger and Cukier, 2013); Big Data (Marr, 2015); Big Data for Small Business for Dummies (Marr, 2016); and Big Data: does size matter? (Harkness, 2016).

Despite the rapid growth in the popularity of big data, pinning down what is meant by the term is more difficult.

Data science is particularly valued for its ability to help predict the future. The desire to predict the future is not new, but while the art of forecasting may have come a long way from divination by chicken bones, it has not necessarily come as far as we might always presume.

The chapter is split into four main parts, narrowing down from predictions in the broadest sense to the practical application for library and information professionals, before looking more closely at statistical methodologies. First it considers the subject of making predictions and forecasts more widely, the importance of recognising the limitations of applying statistical methods to well-defined data sets, and some of the alternative approaches to forecasting. This is followed by an overview of some of the tools and resources that have been used for making predictions with data and the potential application of such methods for library and information professionals. The chapter finishes by looking at the practical application of some specific statistical methodologies: regression analysis and exponential smoothing.

Predictions and forecasts beyond data science

Today's media is filled with an endless stream of forecasts and predictions about what we can expect the future to hold. Alongside our increasingly longrange weather forecasts, our media is saturated with long-term forecasts: central banks make forecasts about inflation and economic growth; health organisations predict the rise or fall of longevity, populations, and the spread of various communicable diseases; and scientists forecast temperature rises and predict the impact of climate change. At the same time there is always an endless stream of pundits ready to predict the doom or utopia just over the horizon for a company, political party or country, based on the various business or political decisions it currently faces. Invariably many of the predictions will be wrong, this is partly the result of conflating forecast and prediction, but mostly because of the complexity of the world.

The terms prediction and forecast are often used interchangeably with reference to anticipated future events, but it can be useful to distinguish the two words by looking at the emphasis they place on existing data. Whereas forecasts are typically more grounded in existing data, predictions may include instances where visions of the future rely on more tangential information.

The aim of this book is not to introduce programming, but rather to introduce data science in a way that shows that it is not difficult to move away from pre-packaged desktop software and start integrating some programming in your own analyses. This appendix includes more detail on some of the programming concepts used throughout this book, and is designed to try and help absolute beginners look more closely at the code rather than gloss over the code they come across in the book.

Like the book as a whole, the appendix is based on the idea that there is a lot that can be achieved with very little code. It is structured according to the main things you need to know to make sense of the programming in this book:

• variables, data types

• importing libraries

• functions and methods

• loops and conditionals.

Each concept is illustrated with examples from Python and R. They are not designed to be comprehensive introductions to the concepts, but rather practical introductions reflecting how they are used within the examples in this book. In the same way that similar ideas can be expressed in different ways in the same language, so it is with programming. When you start programming you will inevitably focus on writing code that works, as you progress however you will want to start writing code that is also efficient and maybe, eventually, aesthetically pleasing.

These examples are available in two notebooks on GitHub, one for the Python examples and one for the R examples: https://github.com/dpstuart/ jupyter/blob/master/Chapter_AppendixPython.ipynb and https://github.com/ dpstuart/jupyter/blob/master/Chapter_AppendixR.ipynb.

Variables, data types and other classes

Programming is about the manipulation of objects represented in a computer's memory. These objects can consist of the inbuilt data types, and newly defined classes of object. The basic data types (e.g. strings, integers and real numbers) are similar across high level languages, however more complex classes of object may be specific to a particular language. A variable's data type also determines which functions or methods can be applied to it.

In essence this book is a natural follow-up to my first book, Facilitating Access to the Web of Data (Facet Publishing, 2011). Many of the ideas the book discussed are now widely accepted: library and information professionals have an increasingly important role in data management and access, data and data science are now the subject of massive and widespread interest, and the software and tools available have developed massively.

Since the first book was written libraries have become ever more involved in data management and access, and whereas the idea that there are advantages in library and information professionals having basic programming skills may once have been controversial, it is now an increasingly mainstream idea, which is widely acknowledged if not universally accepted. Programming skills are complementary skills rather than new skills, pushing the upper limits of how information professionals can help on the spectrum of data analysis. The changing attitude is reflected by including a few code examples throughout the book. It is not necessary for the reader to read or follow all these examples, but information professionals should be aware that there is much more that they can do if they are willing to step away from the comfort of the graphical user interface and engage directly with the data, and the easiest way to do this is by acquiring some basic programming skills. As will be seen throughout this book, in the intervening years it has also become much easier to write and share code.

The importance and potential of data is also more widely recognised than when Facilitating Access to the Web of Data was first published. Data scientist was not labelled the ‘sexiest job of the 21st century’ (Davenport and Patil, 2012) until the following year, when ‘big data’ also burst into public consciousness. Whereas data science was only mentioned once in Facilitating Access to the Web of Data, and big data not at all, unsurprisingly the terms are explored far more thoroughly in this book as librarians are increasingly expected to do more than merely facilitate access to the data that is available, but help analyse, visualise and interpret it.

As has been seen throughout this book, data is not just numbers but also text, and the internet gives access to vast quantities of text online. This varies from the individual keywords that have been entered into search engines or are associated with bibliographic records, to huge quantities of unstructured text in the form of web pages, blog posts, microblogging updates and articles in online databases. This chapter considers some of the ways this data may be analysed for insights.

Following a brief discussion of how text analysis may be applied by library and information professionals, the chapter is broadly split into two halves. The first half considers natural language processing and the analysis of chunks of text, discussing machine learning, sentiment analysis and topic analysis. The second half considers techniques related to keywords or n-grams; it considers term frequency and burst detection. The two parts are not unrelated; for example keywords may have been extracted through natural language processing or created independently as part of a classification process, whether formal (e.g. applying subject terms) or informal (e.g. tagging).

Text analysis and mining, and information professionals

Of the three approaches to data science considered in this book (clustering and social network analysis, predictions and forecasting, and text analysis and mining), it is probably text analysis that has the most widespread potential for library and information professionals. While each of the approaches has its applications in the library, and there is plenty of overlap (e.g. predictive search suggestions), there are numerous potential applications for text analysis in the library because of the pivotal role of text in the history of the library and driven by the huge growth in online content and unstructured data.

The goal of creating a universal library, which contains all the books and useful information ever published, is ultimately unattainable, but great strides have been taken to achieve it. While the digital revolution has reduced the physical barriers to such a library bringing vast reductions in the costs of digitising, copying and storing publications, there are inevitably certain limitations that will never be overcome: history is filled with lost works; copyright holders may object to their works being deposited in such an online library; and there is a variety of grey literature and unpublished manuscripts that are not included in any library's collection policy.

It is well known that we live in an age of unprecedented access to vast quantities of data, and that analysis of this data can produce insights that may be extremely valuable. Less well known is how to analyse the data to gain those insights. That is the topic of data science and this book. This first chapter explores what data science is, some of the drivers behind the rapid increase in data, and how it can be applied within the library and beyond.

By the end of this chapter the reader will be able to see the value of data science beyond the hype, and its widespread applicability within the library and information sector.

Data, information, knowledge, wisdom

Definitions of data science invariably incorporate terms such as data, information or knowledge:

So before we attempt to answer the question ‘What is data science?’, first we must understand what data is, and how it relates to the data–information– knowledge–wisdom (DIKW) hierarchy that is so often discussed by the library and information science community.

The DIKW model, popularised by Ackoff (1989), creates the hierarchy as a pyramid, with each stage building on the one below, with data at the base. In reality, however, the terms are often used with wide ranging and overlapping definitions (Rowley, 2007; Zins, 2007) and such a bottom-up approach fails to reflect the complexity of how information, knowledge and wisdom are actually derived as it relies too much on the data foundations (Frické, 2009). Of course, ‘all models are wrong, but some are useful’ (Box and Draper, 1987), and the fact that the DIKW model is wrong does not mean it is totally without value.

Data science is the process by which insights may be gleaned from the data that is available. This chapter considers some of the different models that have been proposed for the data science process, and provides a five-stage procedure that can be applied by information professionals:

• Frame the problem

• Collect data

• Transform and clean data

• Analyse data

• Visualise and communicate data.

As is discussed below, such a model is necessarily an oversimplification, but nonetheless provides a framework for discussing some of the steps common to many data science studies in the chapters that follow: Chapter 4 considers some of the tools available for each of these stages, before chapters 5, 6 and 7 consider some of the main data analysis techniques more closely.

This chapter starts by discussing the construction of the five-stage procedure, its advantages and (many) limitations.

Modelling the data science process

It is important to recognise that no model captures the myriad of different studies that fall under the banner of ‘data science’. Some start with a question, but others start with the data, the analysis or even a visualisation. So why bother to model the data science process? Primarily to take the complexity of the information ecosystem, which can quickly become overwhelming with its variety of data, tools and questions, and encapsulate the process so that it can be considered more holistically.

There are different ways that the data science process can be modelled. Hayashi (1996), for example, models data science as a three step process: design, collection, analysis. Springboard (2016) breaks down the collection and analysis stages further, reflecting the inevitable need to clean the data that is collected and the different objectives of exploratory and in-depth analysis. Marr (2015) and Springboard (2016) recognise the importance of communicating the findings. Data science is not about idle insights, it is about being able to turn those insights into actions. Table 3.1 summarises these three models of the data science process.

However the process is modelled, it is important to recognise that rather than comprising distinct steps the whole process is iterative, and it is often necessary to return to an earlier step.