In terms of artificial intelligence technologies, large language models are optimised for text (the clue is in the name!), deep learning was primarily aimed at processing images and other media and data analytics deals with numerical data
The term data analytics refers to the science of analysing raw data to make conclusions about information. Many of the techniques and processes of data analytics have been automated into mechanical processes and algorithms that work over raw data for human consumption. Data analytics can be used by different entities, such as businesses, to optimise their performance and maximise their profits. This is done by using software and other tools to gather and analyse raw data.
In practice, data analytics covers a wide spectrum of techniques from simple data analysis to really complex analyses.
Data analytics may be broken down into four basic types:
- Descriptive analytics: This describes what has happened over a given period of time. Has the number of staff working for the business gone up or down? Are sales stronger this month than last?
- Diagnostic analytics: This focuses more on why something happened. It involves more diverse data inputs and a bit of hypothesising. Did the latest pay rise affect staff morale? Did that latest marketing campaign impact sales?
- Predictive analytics: This moves to what is likely to happen in the near term. What happened to staff absenteeism the last time we had a flu outbreak? How many weather models predict a hot summer this year?
- Prescriptive analytics: This suggests a course of action. For example, we should offer free flu vaccinations to staff if the likelihood of a serious flu outbreak is above 50%,
Data analytics underpins many quality control systems in the business world. It’s nearly impossible to optimise something if you aren’t properly measuring it, whether it’s your weight or the number of staff days lost to ill-health.
Data analysts can use several analytical methods and techniques to process data and extract information. Some of the most popular methods include:
- Time Series Analysis: Tracks data over time and solidifies the relationship between the value of a data point and the occurrence of the data point. This data analysis technique is usually used to spot cyclical trends or to project financial forecasts.
- Regression Analysis: This entails analysing the relationship between one or more independent variables and a dependent variable. The independent variables are used to explain the dependent variable, showing how changes in the independent variables influence the dependent variable.
- Factor Analysis: This entails taking a complex dataset with many variables and reducing the variables to a small number. The goal of this manoeuvre is to attempt to discover hidden trends that would otherwise have been more difficult to see.
- Cohort Analysis: This is the process of breaking a data set into groups of similar data, often into a customer demographic. This allows data analysts and other users of data analytics to further dive into the numbers relating to a specific subset of data.
- Monte Carlo Simulations: Models the probability of different outcomes happening. They’re often used for risk mitigation and loss prevention. These simulations incorporate multiple values and variables and often have greater forecasting capabilities than other data analytics approaches.
In the same way as the massive and growing amount of data has facilitated the training and development of large language models, so it has enabled the development of a type of data analytics known as Big Data. At its most basic, big data analytics is the process of examining large, complex datasets to extract valuable insights and knowledge.
The Big in Big data doesn’t just refer to the quantity of data but to five characteristics often referred to as the “5 V’s”: volume, velocity, variety, veracity, and value.
- Volume: The sheer amount of data is vast.
- Velocity: Data is generated and processed at high speeds.
- Variety: Data comes in various formats (structured, semi-structured, unstructured).
- Veracity: The accuracy and reliability of the data are important.
- Value: The insights derived from the data must be meaningful and actionable.
Together they correspond to a major increase in the complexity of data analytics. At the same time, the sheer quantity of data available provides new opportunities analogous to the training sets for large language models.
Big data analytics uses a range of tools and techniques, including machine learning algorithms that allow computers to learn from data, data mining to discover patterns and relationships in data, traditional statistical analysis methods to analyse data and data visualisation to create visual representations of data to illustrate what is happening.
Big Data Analytics is designed to improve decision-making by uncovering hidden insights, so that organisations can make more informed data-driven decisions. This in turn, should increase efficiency by helping to identify areas for improvement and streamline processes.
Particularly in the retail sector, big data is credited with enhancing business understanding of customer behaviours, and this has allowed businesses to personalise their offerings and improve customer satisfaction.
Businesses that effectively leverage big data analytics can gain a competitive edge by making better decisions, responding faster to market changes, and reducing costs by identifying inefficiencies and optimising operations.