2.3 Tasks
2.3.4 Perform Data Analysis
Guide to Business Data Analytics
Data analysis involves the extensive deep analysis performed once the data quality issues are resolved through exploratory analysis. Performing data analysis involves the application of mathematics, statistics, and the completion of extensive mathematical analyses related to answering the research questions for different stakeholders.
Where exploratory analysis tested the dataset, performing data analysis involves using the results of an exploratory analysis to determine the best mathematical methods and approaches to use and then conducting the in- depth data analysis required to answer the analytics problem. The original research question in the business language is transformed into a mathematical question, which is translated into a model to perform a deeper analysis.
When performing data analysis, data scientists use technical techniques requiring extensive mathematical skills. Some techniques are leveraged to find associations or to cluster data, which is helpful when identifying patterns (for example, association rule learning, decision tree analysis, and K-means clustering.) Many techniques, such as the use of machine learning and artificial intelligence, advance the data scientist's analysis capabilities. Data scientists may use regression analysis to predict and forecast. Simulation can be used to play out a series of actions or behaviours.
Many of the algorithmic models are automated through available machine learning packages, but the parameters and measures of success require significant business domain knowledge to be aligned to the research questions. For example, an organization seeking to discover market segments based on the order history and customer profile data may require some marketing guidance on how the segments should be discovered. It may depend on demography, monetary value, type of products the customers purchase, customer lifetime value, and other factors. The choice of the clustering model may be influenced based on the business direction. Similarly, most of the success criteria for models are also influenced by business decisions that need to be translated into mathematical parameters where analysts may contribute significantly by describing the success criteria to the data scientists.
Data scientists use creative thinking skills to determine different approaches for answering the research question, especially when the data results are not helping to achieve the stated objectives. As with exploratory analysis, the data scientist uses industry and business domain knowledge, and when not present, these skills can be augmented by leveraging the skills of the business analysis professional.
Where exploratory analysis tested the dataset, performing data analysis involves using the results of an exploratory analysis to determine the best mathematical methods and approaches to use and then conducting the in- depth data analysis required to answer the analytics problem. The original research question in the business language is transformed into a mathematical question, which is translated into a model to perform a deeper analysis.
When performing data analysis, data scientists use technical techniques requiring extensive mathematical skills. Some techniques are leveraged to find associations or to cluster data, which is helpful when identifying patterns (for example, association rule learning, decision tree analysis, and K-means clustering.) Many techniques, such as the use of machine learning and artificial intelligence, advance the data scientist's analysis capabilities. Data scientists may use regression analysis to predict and forecast. Simulation can be used to play out a series of actions or behaviours.
Many of the algorithmic models are automated through available machine learning packages, but the parameters and measures of success require significant business domain knowledge to be aligned to the research questions. For example, an organization seeking to discover market segments based on the order history and customer profile data may require some marketing guidance on how the segments should be discovered. It may depend on demography, monetary value, type of products the customers purchase, customer lifetime value, and other factors. The choice of the clustering model may be influenced based on the business direction. Similarly, most of the success criteria for models are also influenced by business decisions that need to be translated into mathematical parameters where analysts may contribute significantly by describing the success criteria to the data scientists.
Data scientists use creative thinking skills to determine different approaches for answering the research question, especially when the data results are not helping to achieve the stated objectives. As with exploratory analysis, the data scientist uses industry and business domain knowledge, and when not present, these skills can be augmented by leveraging the skills of the business analysis professional.
