Business Analytics

A Contemporary Approach

Thomas W. Jackson, Steven Lockwood

• Publisher: Bloomsbury Academic
• Year: 2018
• Print ISBN:
  9781137610607
• Edition: 1
• eBook ISBN:
  9781350304765
• Topic: Business
• Subtopic: IT Industry
• Index: Business

This innovative new textbook, co-authored by an established academic and a leading practitioner, is the first to bring together issues of cloud computing, business intelligence and big data analytics in order to explore how organisations use cloud technology to analyse data and make decisions. In addition to offering an up-to-date exploration of key issues relating to data privacy and ethics, information governance, and the future of analytics, the text describes the options available in deploying analytic solutions to the cloud and draws on real-world, international examples from companies such as Rolls Royce, Lego, Volkswagen and Samsung. Combining academic and practitioner perspectives that are crucial to the understanding of this growing field, Business Analytics acts an ideal core text for undergraduate, postgraduate and MBA modules on Big Data, Business and Data Analytics, and Business Intelligence, as well as functioning as a supplementary text for modules in Marketing Analytics. The book is also an invaluable resource for practitioners and will quickly enable the next generation of 'Information Builders' within organisations to understand innovative cloud based-analytic solutions.

Part I

CONTEXT

CHAPTERS

1 UNDERSTANDING THE BIG DATA LANDSCAPE

2 ANALYTICS: DESCRIPTIVE, PREDICTIVE, PRESCRIPTIVE AND COGNITIVE

1

UNDERSTANDING THE BIG DATA LANDSCAPE

1.1 Introduction

The speed of life seems greater than ever. Nowadays we are aided by the digital age, where the latest gadgets and gizmos keep us in touch with what is going on 24/7. Thanks to this, we are probably the best-informed generation ever. However, with digital technology there is a continuous battle to try to harness it and make it work for us rather than allowing it to dictate our daily lives, as is so often the case. With so much data and information being generated – for example, every 60 seconds, 100 hours of YouTube video is uploaded, 433k tweets sent and 204 million emails sent, we also face the battle of how to cope with increasing volumes of unstructured data and what to do with it. The effective management and efficient use of data is a major challenge for government, academia and industry. The scale, diversity and distributed nature of current and emerging data assets are increasing, for example through the realisation of the Internet of Things (IoT) through landmark reports like Industrie 4.0, which we will cover in this chapter. What is clear from these big data developments is that the lines between social and enterprise data have become fuzzy. But has the possibility of providing better insights into how to make an organisation more successful been fully realised? Consider healthcare provisions, for example. Think about the Fitbits that many wear and which could connect you to your doctor for 24/7 monitoring. There is the potential for your doctor to provide early intervention to improve your lifestyle, which could change your life insurance premiums. Of course, the next step would be to remove the doctor from the process altogether, and it could all be controlled by the various analytic techniques discussed in Chapter 2. Clearly, we are not at this point yet!

These are very exciting times to be involved in big data and this chapter sets out the big data landscape by covering the very basic concepts of the building blocks of big data through to discussing the exciting new developments of the big data world that will change our lives.

1.2 World Wide Web underpins everything we do

Without the World Wide Web (WWW) none of the changes we see in our lives could have happened. When first used, the Web was seen as no more than a vast search engine, where users could find all manner of useful, or indeed useless, information. This is still a crucial aspect of the Web but other interesting uses for it were beginning to be identified and exploited around 15 to 20 years ago when all aspects of traditional services began to change under its influence.

The first real large-scale change was in the use of online banking and shopping. The idea for online shopping had been around since the late 1970s when an Englishman named Michael Aldrich invented videotext, a modified domestic TV that connected to other systems using a telephone line and modem. Even though crude in our thinking today, the WWW introduced a level of scale not previously seen and companies such as Amazon and eBay (originally known as AuctionWeb) were established and grew rapidly. In the UK, traditional outlets such as Argos and Screwfix also created online shopping experiences, as did the world’s largest retailer Walmart, with many others swiftly following suit. Even shopping for holidays changed dramatically, from going to the travel agents to book a holiday, to shopping online for the best deal. For Christmas 2016 the estimates for sales over the period leading up to Christmas through online sales alone was around £126 billion. We are rapidly reaching a point where most sales will be online. These scenarios are driven by structured information, such as stock, sales and prices.

The next big change the WWW enabled was the growth in social interaction. Blogs, chatrooms, wikis, social websites such as Facebook and more recently Twitter, Snapchat and Instagram have all arrived. These scenarios are using unstructured data (text, video, audio, images). This has allowed for individuals, groups and things (think chatbot!) to be constantly digitally connected anywhere across the globe where the WWW can function. We can now extend our reach to a far wider audience than ever before.

The third big change is the use of the WWW as the medium to move sensor data around. We can now connect almost anything to the cloud and operate it through the cloud (Rittinghouse and Ransome, 2016). This creates the opportunity to be able to understand just about anything we wish to, whether it be migratory paths of geese, the environmental impact of pollution, the movement of goods across a factory floor or an individual’s basic health characteristics (think about all the data obtained through something like a Fitbit – steps walked, distance travelled, calories burnt, sleep quality – and which could be merged with information from your smartphone – your location for example). Sensors can be attached/embedded in all these scenarios and a wealth of useful data obtained for analysis to understand and hopefully improve things (Seth, 2016). The interesting point here is that when we start to bring all this data together we can use analytics to identify trends, predict future patterns, identify best- and worst-case scenarios, and suggest next best actions to individuals based on similar peer-clustered data.

Many of these changes have been inspired by business or social pressures using the technologies we have had around for the last 20 years or so, gradually increasing the scale and complexity of those solutions. For example, the rise of internet speeds has continued at pace, enabling businesses and citizens to partake in activities not feasible even five to ten years ago – streaming video and content in general is an obvious example here, but the general movement of large quantities of data is not feasible without broadband. Figure 1.1, from the ISPreview website, shows the top 10 global average connection speeds. It’s worth noting the UK and the USA are not in that list. Browsing is no longer just a PC/laptop process; Wi-Fi hotspots, 4G and 5G mean we can go online just about anywhere. This ability is important in changing how we have evolved to use the WWW.

Figure 1.1 Global average broadband speeds

Source: Copyright © ISPreview.co.uk. Reproduced with permission.

1.3 The building blocks of big data

The contemporary literature views the growth in big data as directly attributed to accessibility of technology and new forms of distributed data storage. Hashem et al. (2015) characterises a common view, attributing the growth of big data to ‘social media, IoT and multimedia’ (2015, p. 99). Both Chen (2014) and Hashem (2015) warn of a consistently greater exponential increase of data from the IoT, with Jara et al. suggesting ‘unprecedented scenarios of interaction’ (2014, p. 379).

In recent times the world has witnessed a dramatic increase in the amount of formal and informal information produced in a crowded ‘information society’, in which the amount of information being created every two days is equivalent to that which had been created from the dawn of civilisation until the year 2003 (Seigler, 2003). This means that one of the fastest growing quantities on this planet is the amount of data/information being produced. We could frame it as Clive Humby does, ‘data is the new oil’, and as with oil there are dangers ahead, as we now know that oil has proven to be a polluter. We need, however, to understand the building blocks underpinning big data, which can be broken down into five key categories: Data, Information, Knowledge, Understanding and Wisdom. These are represented by Figure 1.2, which shows the size of the categories and the level of difficulty to reach the top, wisdom. It is helpful to provide a definition of these five categories as founded by Russell Ackoff, a systems theorist and professor of organisational change (Ackoff, 1967).

Data – In simple terms, characters which are symbols, numbers or letters that we might interact with but which have no context. An example of data is a spreadsheet that has no row or column information which provides the context on how to interpret the data.

Figure 1.2 The building blocks – from data to wisdom

Information – Data that is processed to aid decision-making, i.e. provides answers to ‘who’, ‘what’, ‘where’ and ‘when’ questions. Therefore, once the context becomes sufficient to gain understanding of what the data is, we have information. For example, a spreadsheet that has data 090917 in one column and 10 in another means very little. However, if you add context of ‘date’ and ‘temperature °C’ you are provided with information that is usable, especially if the filename is titled ‘Sydney Weather 2017’.

Knowledge – Application of data and information; answers ‘how’ questions. Knowledge is created by processing information and then applying it to create a feedback loop to further one’s knowledge about the information. For example, if you lived in Sydney, Australia you would know that in September if the temperature was 10⁰C then that would be very cold. You would then question if the data was wrong or had global warming taken effect (which moves into Wisdom).

Understanding – Cognitive comprehension of ‘why’. Understanding: the time that knowledge can be reasoned. It is the process by which one can take knowledge and synthesise new knowledge from the previously held knowledge. So this is more than memorising information; it is about learning. So if we knew the temperature in Sydney should be around 20⁰C in September and we know about climate change then we may have a better understanding of why it is so cold. Historically, machines have struggled to make this leap but deep learning is fast encroaching on this space and in the cognitive computing domain we are likely to see big developments as we approach and then transition into the next decade.

Wisdom – Evaluation of understanding. Unlike the four categories above, it’s the ability to think about a question to which there is no (easily achievable) answer, and in some cases, to which there can be no humanly known answer, period. Wisdom is about our mental ability to determine – based on our knowledge and understanding – if something is a good or bad idea. For example, in the case of an autonomous car, what decision will it make when given the scenario the car has to crash and it can either injure the passenger or run over the pedestrian. This involves snap choices, ethics and morals which we will discuss in Chapter 8.

According to Ackoff, the first four categories relate to the past as they deal with what has happened or what is known. It is only the last category, Wisdom, that deals with the future as it incorporates vision and architecture. The ‘Wisdom’ category has the most value to society, and through the introduction of data analytics, organisations and wider society are able to understand why things have happened and in some cases predict what will happen next. Of course, there have been mixed results, and a good example of this is weather forecasting (meteorology) which dates back to 3000 BC. Although techniques have become much more accurate, the forecasts are still a hotly debated topic because they are still inaccurate, despite having improved significantly in the last 20 years. For example, predicting the minimum temperature on the first night of a forecast has improved over the years to the point where 86.5% of minimum temperature forecasts are accurate to within +/−2°C, over a 36-month average. Not surprising, then, that in 2017 a four-day forecast is more accurate than a one-day forecast in 1980 (Met Office, 2017).
With the building blocks in place, let’s take a look at breaking down the Data category to aid us in understanding the different elements that need to be taken into consideration when dealing with big data. A popular classification of big data is defined by Laney (2001): Volume, Velocity and Variety, as shown in Figure 1.3. These three categories denote the size, speed and heterogeneous structures common across big data. Several definitions have since been applied to Laney’s original classification, such as Veracity (Chen and Zhang, 2014), and more recently the cost or Value from the insights gained. So what do the 5 Vs mean and how do they interconnect?

Variety denotes the range and different formats of data available for use in today’s world. Data today looks very different from data of the past. The enterprise needs of today are to manage this data as is, in its original format, or ...

Table of contents

• Cover
• Title Page
• Copyright
• Contents
• List of Figures
• Foreword
• Preface
• Acknowledgments
• About the Authors
• PART I CONTEXT
• PART II ARCHITECTURAL CONSIDERATIONS
• PART III PRACTICAL APPLICATION ISSUES
• PART IV FUTURE DIRECTIONS
• Appendix 1: Big Data Analytics Platform – Additional Components
• Appendix 2: Graph Databases
• Appendix 3: Open Standards
• Glossary
• Index