Back to glossary

Data Analytics

Also known as: Data analysis, Big Data Analytics, Online Analytical Processing (OLAP)

Data analytics is the qualitative and quantitative processes around collecting, managing and analyzing raw data to generate correlations, patterns and useful meaning that help optimize productivity, gain competitive business advantage and guide strategic decision-making. Data that is extracted, categorized, computed and organized typically come from a large variety of sources, and the techniques used depend on business or organizational requirements. This includes insights into past performance, as well as predictive and prescriptive explanations. It can collect, compare and measure both numerical data for statistical measurement and analysis, and interpret and explain non-numerical data such as text, audio, images and video.

The data analytics of today largely involves automated, mechanical processes and systems that use machine algorithms and artificial intelligence (AI) to extract, transform, model and reveal critical anomalies, changes and metrics that can be lost in mass volumes of information. Skilled specialists who are versed in statistics and can perform self-service analysis in addition to the many streamlined techniques available for the regular business user are called data scientists.


The many types of data analytics

Data analytics is a pervasive term used to describe all forms of processing that examine historical data to draw conclusions and support action. The main components include:

Descriptive analytics

The foundational stage of the business intelligence and statistical data process that aims to collect, organize, model and summarize raw historical data for high-level analysis of critical metrics and useful information. Descriptive analytics is all about answering what happened after the fact with easy to understand data visualizations and data aggregation rather than complex forecasting, and is best suited for day-to-day operations. Examples include data warehousing, inventory management, revenue reporting and marketing performance dashboards. These methods often help prepare the data for further necessary analysis.

Statistical analytics

A more scientific approach at collecting and analyzing every single data sample in a population to identify changes, patterns and trends using raw numbers to eliminate potential bias and modelling to summarize the nature, relation and validity of the model to users. Statistical analysis can also include predictive techniques to anticipate future performance.

Predictive analytics

A subset that uses emerging technologies such as artificial intelligence, data mining, deep learning, forecasting, machine learning and statistical modelling to analyze business data and provide predictions on immediate and far future events. Organizations typically combine predictive analytics with historical data to build a mathematical model, which their analytics software can then use to predict future events based on identified trends in their data. The results act as a guide the business can base decisions on to drive desired outcomes.

Prescriptive analytics

A specialized stage that analyzes data to factor the best course of action for a business or user from several potential scenarios or situations. Recommendations are based on historical, current and future performance, supporting decisions from immediate to long-term. It uses advanced techniques such as AI and machine learning to process high volumes of data and uses the estimations provided by predictive analytics systems on what is likely to happen to then generate answers on the next stage – the best actions to take now.

Automated analytics

Automated analytics is the process of automatically monitoring, detecting, analyzing and alerting a business to relevant anomalies, patterns and trends in data using techniques such as machine algorithms and artificial intelligence to deliver insights to business users instantly, with minimal manual user-analysis or IT intervention necessary. It can help organizations track user-defined metrics and discover critical factors that align with desired business outcomes that may not be known or easily possible through manual analysis. Findings are delivered at fixed intervals or triggers, and monitoring and analysis is done so continuously.

Augmented analytics

Augmented analytics is an advanced process of analysis that leverages AI, machine learning and natural language generation and querying to transform how analytics are built, consumed and shared throughout a business. It streamlines previously specialized data science and machine learning (DSML) tools to make the data preparation, discovery and insight generation process accessible and usable for regular business users, analysts and specialists alike.

Embedded analytics

Embedded analytics is the integration of analytics solutions such as dashboards, reporting and data visualizations into software applications or transactional process systems. 

Contextual analytics

Contextual analytics is a specialized and mature form of embedded analytics that directly integrates analytical components like alerts, charts, visualizations and tables into an application. It focuses on seamlessly blending these elements into the user interface and transactional workflow of an application to make relevant data available at the exact point it is needed in the context of the overall application workflow. 


The best method of data analytics

Ultimately, the type of data analytics your organization or users may choose is dependent on several criteria, including but not limited to the business use case, internal skill-set, desired outcomes, data sources and much more. Data analytics as a whole is a vast and ever-evolving field, but there are thankfully modern analytics platforms such as Yellowfin provide many useful applications and solutions which can suit a variety of analytical needs today.