A brief introduction to Data Analytics
Every day we generate a significant amount of data from our credit card transaction, emails, social media platforms, online shopping and any other actions that require the use of the internet and smart devices. It is reported that we create approximately 2.5 million terabytes of data each day and by the end of 2025, the number will increase to 463 exabytes (10006 bytes) of data in a day! The impact of the pandemic is likely to increase the forecasted number even more, due to the rapid digitalization of organizations and the surge of internet usage by individuals.
The new data economy and digital innovation are reshaping the way most industries function today. Many companies have started to recognize the significance of data collection and data analytics, and have begun to implement in every aspect of their business operations. By using the right techniques and tools, the data generated can be used to gain insights for better decision making. The method to do so is known as data analytics.
Data analytics is a discipline focused on extracting insight from data which include the process of analysis, collection, organization and storage of data. In general data analytics can be categorized into 4 types which are the Descriptive Analytics, Diagnostic Analytics, Predictive Analytics and Prescriptive Analytics. The definitions below explaines the different categories in the simplest way.
The organization abilities in data analytics must follow the sequential process and it starts from the 1st level which is descriptive analytics. It is not possible to progress to the higher level without this in place. Descriptive analytics can show us what has happened based on past data. It is the most fundamental form of data analytics that is being used in every organization to show the usual results such as sales revenue per month, sales margin of products and revenue for each country and others. Based on historical data It helps to convey results and insights in the form of data visualizations like charts, graphs, reports and dashboards.
Diagnostic analytics is the use of techniques to find out why did something happen? For example, why was there a significant rise in sales during certain months or why was there a decrease in the number of website visits? In comparison, diagnostic analytics is a harder process than descriptive analytics as it requires deep-dive into the data to search for insights and uncover the reasoning behind the results presented. The different techniques can include data discovery, drill-down, data mining and correlations. Unlike descriptive analytics, where in general can be done by almost everyone with basic knowledge of Excel, diagnostic analytics may require an individual with the right knowledge to get the job done.
Predictive analytics is forward-thinking, it is the way of using data gathered to predict the future. The basic of predictive analytics is modelling. Analysts build models to best represent the real-world environment and use the models to simulate the behaviour of the system and to predict the potential outcomes in the future. There are many applications where predictive analytics can be applied such as predicting consumer buying behaviour, forecasting inventory requirements or even complicated task such as predicting the travel path of an aeroplane. As you can imagine, it is a more complex process than previous types or analytics. There are different predictive modelling techniques such as decision trees, regression techniques and neural networks.
Decision trees are capable of demonstrating statistical probability by showing every possible outcome of a particular decision and how one outcome or choice can lead to the next. Regression techniques on the other hand help users to understand the relationship between different variables. And lastly, the neural network uses algorithms to identify underlying relationships within a data-set mimicking how the human minds work. Machine learning is also frequently used in predictive analytics, especially in a situation where we do not know the appropriate models to define a particular system. Machine learning uses data to build models on its own and put them to test in an iterative manner and readjust as it learns from the outcome of each test and eventually arriving at the most accurate model.
Prescriptive analytics is also forward-thinking and it builds upon predictive analytics. The main difference is that we use a machine to make recommendations for us. The machine or system uses data to predict and recommend the best possible action. The most relatable example of all would be our navigation software application such as Waze. It is capable of interpreting data such as traffic condition, weather forecast and other factors to recommend the best or shortest route for us to reach the desired destination. Prescriptive analytics requires the support of different technologies such as artificial intelligence, machine learning and Big Data.
If you would like to learn more about data analytics and big data, sign up for our weekly newsletter now or visit our knowledge centre. For more information about our data analytics training or consulting services please contact us at email@example.com.
Leave A Comment