3 Data everywhere

Thinking back to household smart devices, you might wonder: How does a streaming app know what to recommend to you? How does my robot cleaner know where to clean? Well, data plays a crucial role. These devices rely on data to operate efficiently. Data is collected to make your experience more personalised and automated.

Data is information that is collected, stored, and used for various purposes. It can come in many forms, such as text, images, and numbers. AI utilises data for learning and reasoning. For example, when you consistently watch shows of the same genre on a streaming app, this data helps the AI model determine what to recommend to you next.

In the first video, you’ll gain an understanding of what data is and why it’s essential for advancing AI.

Download this video clip.Video player: Video 4

Show transcript | Hide transcript

Transcript: Video 4

Voice Over:

Have you ever wondered why data collecting is so important in today’s world and what it can be used for?

Well, data is the raw material AI systems use to analyse and make predictions, and you probably produce way more data than you realise, even just through your mobile phone.

Back in the 1990s, the rise of the internet marked a significant leap in access to data and the use of AI. This era set the stage for the data-driven world we live in today.

Fast forward to the present, we have no shortage of data fuelling analyses and personal pattern recognition processes.

A man looks at a laptop.

Just imagine how much data is being created while you watch this video.

Maybe you are sending a text message to a friend. At the same time, someone next to you might be on a call.

Somebody else has just posted a photo on their social media page while another has started a new video series that piqued their interest on a streaming app.

Correct... all these actions create new data.

Data is constantly being generated with each passing second.

Yet this raw data requires careful refinement.

Data needs to be refined because the quality of the dataset used to train algorithms have a direct influence on the accuracy of the AI model.

That means that diverse and representative high-quality data is essential to developing advanced AI models.

Consider a music streaming service. It uses the data from your listening habits, such as the songs you skip, the ones you play on repeat, and the playlists you create, to recommend new music that you might like.

This is a practical example of how data is used to enhance our everyday experiences and how datasets train the AI models to recommend the next song for you.

And that is what we use machine learning for, to analyse and learn from the data. But it does not stop there.

Once we have this information, we can use it to train AI models, improve products and services, make predictions, and even uncover new insights that were previously hidden.

So, you see, understanding data is not just about quantity, but also about quality and diversity.

It is about recognising patterns and making connections that might not be immediately obvious. It is about using these patterns to predict future trends, behaviours and outcomes.

So, the next time you use your mobile phone, remember that every action you take, every piece of data you generate, contributes to this vast, interconnected web of information.

And it is through understanding this data that we can truly unlock the potential of AI.

End transcript: Video 4

Download

Video 4

Data allows AI to learn, adapt, and make informed decisions. However, it isn’t just about the quantity of data available but also the quality. The effectiveness of AI models is directly tied to the quality of the training data.

Activity 2 Patterns in data

Timing: Allow approximately 10 minutes

In this video, you’ll discover why finding patterns in data is important and how it can enhance user experience. You’ll also learn about the impact of AI on our daily routines.

How can finding patterns in data help a company improve its service for users? You can type your answer in the space provided below.

Download this video clip.Video player: Video 5

Show transcript | Hide transcript

Transcript: Video 5

Voice Over:

Data is constantly being created and refined to train AI models.

But what exactly are these models looking for in the data?

Patterns are identifiable repetitive behaviours.

AI models are particularly good at recognising these patterns, mainly because they have access to a large volume of data.

Let’s consider LinkedIn as an example. Have you ever noticed how LinkedIn suggests people you might know or jobs that you might be interested in?

It is not just based on your connections or job search history, but also on the behaviours of people all around the world who have similar connections or job interests.

What is really interesting is platforms like LinkedIn may look for patterns not only within your professional network and job preferences, but also in relation to the specific times you are active on the platform.

For example, many people tend to browse apps during their lunch breaks or after work hours, so apps may suggest new connections or posts at those times too.

This pattern identification does not just occur on platforms like LinkedIn, but with many companies across many apps and platforms.

Imagine the benefits of understanding what users really want or even creating new consumption trends based on the market. This understanding allows companies to customise their services for each user, thereby enhancing the user experience and boosting satisfaction.

It is about using data to meet the needs of the user in the most effective way possible.

Think about the World Cup. What are potentially the best-selling products before the tournament?

You might be thinking team shirts.

Recognising this trend of team shirt consumption can provide a unique competitive edge for companies. It could even help in the new development of products or features.

Every click, every like, every search - these are all pieces of data. AI models process this vast amount of data to notice patterns.

These patterns help the AI model tailor your experience, making it more relevant and engaging.

It is like having a personal assistant who knows exactly what you need, even before you do!

But how does an AI model learn to recognise these patterns?

Well, it does so through machine learning.

End transcript: Video 5

Download

Video 5

Interactive feature not available in single page view (see it in standard view).

To use this interactive functionality a free OU account is required. Sign in or register.

Interactive feature not available in single page view (see it in standard view).

Discussion

Finding patterns in data helps companies improve their service by enabling them to analyse, understand and predict customers’ behaviour at scale. AI can analyse patterns across millions of users rather than guessing them. When activity spikes at certain times, what posts get shared, or which items sell well, become clues for smarter business interactions. These hints let services adjust on a deeper level, quietly tuning the way customers engage. For example, by recognising that users with similar professional backgrounds engage with certain job postings, or that sports fans purchase team merchandise when major sporting events approach, companies can deliver timely, relevant recommendations that match the needs of their customers. This pattern recognition transforms raw data from simple, interactions such as searches and clicks, into valuable insights, allowing companies to customise their offers and develop new features based on emerging trends to drive profits.

AI models use data to identify patterns. At the core of this process are algorithms that analyse data fields, learning from the patterns within the data to generate models. These models are then used to make predictions or decisions about new data. This process is called machine learning.

The quality and quantity of the data are crucial. High-quality data ensures that the AI model can learn accurately and make reliable predictions. Poor-quality data can lead to incorrect outcomes. Therefore, data preprocessing, which includes cleaning and organising data, is an essential step in the machine learning process.