The digital world is full of new and changing jargon, and it's important that savvy business leaders understand at least the basics of these so they know what they're investing in and why it's important.
Today we're going to cover five key terms:
What do they mean, how do they work, and how could they apply to businesses?
Data science is the point where statistics, artificial intelligence (AI), analytics and the scientific method intersect. People who work in data science are often called data scientists, and their role is to parse large datasets, combine different types of data (i.e. customer, sales, inventory) and figure out both problems and solutions.
Common sources of data
AI can also come into play here. Artificially intelligent algorithms, especially those powered by machine learning, are able to automate much of the data gathering and visualization process - making the lives of a data scientist much easier, and their work much quicker. We'll talk about AI in more detail later.
Data scientists are scientists. They use the scientific method to research and understand a problem, then find empirical evidence to draw fact-based conclusions. Data is, therefore, fundamental to their work, and so their scope includes the gathering, preparation and maintenance of data sets, as well as figuring out how to use them.
They are also often skilled programmers. Data scientists will often write and/or use tools to:
So why is this important in a business context?
The goal here is to derive insight from data to drive strategic decision making - making decisions based on hard fact, not gut feel.
If a company can make these calls using evidence, they are often better able to grow their business or make changes in a cost effective and strategic manner. New ventures are less likely to fail.
Data science also generates more than analytical insights. For example, it can be used to build self-learning programmes that utilize data to improve process or human inefficiencies. Therefore, data science is vital for the smooth functioning of AI and machine learning, as defined below.
Is this different to data analytics?
Data analytics is a subset of data science - let's cover that separately next.
Learn more: "Why processes, not technology, are key to unlocking data"
Data analytics is the gathering and analysis of large sets of data. Its purpose is to draw meaningful insight from large quantities of information - finding answers to those important questions.
Much of this process can be automated, meaning even less-experienced individuals can start to analyze data and gain reasonably meaningful insights with simple beginners tools.
Data analytics: This is all about the gathering and utilization of data to inform. Analysts design and maintain databases, ensure coherence between datasets, and create visualizations that can be communicated to others.
Data science: This can often mean the same thing, but the scope is more broad (the former being a subset of the latter). Whereas analytics is about finding answers, data science could be described as finding the right questions to ask.
Example: A data analyst is asked to find ways to improve, say, the efficiency of a retail store. A data scientist is the one that, through the use of vast datasets and different sources as well as machine learning algorithms and other tools, will tell you that those inefficiencies exist in the first place.
People often classify data analytics with different terms, but ultimately they tend to fall into these four buckets:
Artificial intelligence, or AI, is an umbrella term for a broad set of technologies designed to perform tasks and understand inputs on their own. Of course, a lot of scientists want to create artificial human intelligence, but modern AI is much more simple.
AI isn't just another word for automation, it's smarter than that. As described in this article, AI generally has three components: Intentionality, intelligence and adaptability. It can receive inputs and learn from them, improving its own efficiency. This is called machine learning, or deep learning (see below). AI is exceedingly accurate and growing increasingly intelligent.
AI is key to unlocking the next stage of technology. It's not typically a tool offered by itself (as in, you don't generally just buy an 'AI'), but rather it is woven into the fabric of another tool designed to solve a particular problem - for example, self-driving cars, self-monitoring buildings, healthcare tools that can diagnose conditions.
But, AI may also be more humble than that. It could be a smart weather app, a digital personal assistant with basic conversation functionality, a chatbot for a website, even a simple spreadsheet that is capable of filling itself out in more complex ways than just copying and pasting.
Learn more: "How to implement digital transformation"
Machine learning is a subset of AI to do with self-improvement. It is a combination of algorithms, data and training that, together, give computers the power to improve themselves. Machine learning can, therefore, take in inputs, spot patterns and design its own efficiencies.
This technology is built to go beyond human capability. Not in a 'take over the world' sense, but in an efficiency and scalability sense. Machine learning can process vastly more data than a human ever could in far less time, then build patterns that humans might not see.
Alongside machine learning, you may also hear the common terms 'deep learning' and 'neural networks'. So, are they relevant?
Defining neural networks: A neural network is a mathematical model that uses computer algorithms designed to mimic the human brain - that is, an interconnected set of 'neurons' which can each receive an input and pass on information or a command. This is a popular way to design a machine learning program, as neural networks have proven effective at understanding highly difficult problems - like speech or image recognition.
Defining deep learning: Deep learning is a way to describe a particular type of neural network, that is, a 'deep' one. At its simplest explanation, the 'depth' of a neural network refers to the number of nodes and streams in the model. Each node is like a filter for information, processing and asking questions before passing on the input to another node. The more nodes, the more the model is able to compute data with increasingly complex filters. To use an example, imagine asking a network of very few node streams if an image contains a cat. Our shallow network here can only make a few decisions - "Does it have eyes? Does it have fur? Yes, it's a cat." But, with more nodes come more questions, allowing for a more precise answer.
While AI is key to unlocking the next stage of technology, machine learning is key to unlocking the next stage of AI. It takes software from simple, repetitive automation to a genuine intelligence. If something can receive inputs and learn from them, it is going to be far more cost effective and produce far better results over time.
Additionally, machine learning - like AI - can be applied in both large and small capacities. Self-driving cars are powered by highly complex learning algorithms, but your local food delivery app can also learn (like what kind of foods you like) so it can improve its service.
If AI is the brain, robotics is the body.
Robots are machines that can be used to either replicate or improve a human-performed task. Think production lines in car factories, surgery robots for remote surgeries, bomb disposal robots, care robots in rest homes, robotic waiters … these days the list is endless!
But robots don't have to be smart. They can also be very simple, repeating one process over and over. These machines don't learn from the experience, but can produce highly accurate, fast, efficient results with limited or no variation.
Robots are commonly used in sectors such as manufacturing because of their precision and efficiency. They can produce a lot of products with very few variations while working tirelessly. But, their use is expanding into other sectors. Self-driving cars are, again, a great example. Care home workers, too. Of course, there's also the Google bot that defeated a world champion at Go.
Robotics can also be used to keep people safe. It's dangerous for humans to enter certain hazardous situations, for example, but robots can be sent in to assess or control that same situation. Hazards could include bomb threats, natural disasters or chemical spillages.
Will robots replace humans?
Some people believe robots will replace humans, and while there have been some instances of replacement around the world, typically robots are used in a human enablement capacity. Robots handle the menial work so that their human coworkers can focus on value-adding activities, strategy, or just generally more interesting jobs.
The debate on robots replacing humans is ongoing, and will continue for many years more.
At dig8ital, we've been helping organizations around Europe to digitally transform for years, backed by our team of industry experts. We can work with you to understand where your business is, where it needs to go, and the steps it must take to successfully join one with the other - from finding skills gaps to immature processes, protecting against cyber crime, and investing in digital technology like AI.
To learn more about how to optimize your business for digital transformation, download our PDF here. Otherwise, talk to use today for a free maturity consultation and we'll help figure out your unique needs.