Data Analysis in Power BI

You’ll begin this Exploratory Data Analysis (EDA) course by learning how to use descriptive statistics, identify missing data, and apply imputation techniques to fill the gaps in your data. You’ll also build histograms to analyze distributions and use winsorizing to remove outliers.

Now it’s time to explore the relationships between categorical variables using proportions. You’ll then use box plots and descriptive statistics to determine how a continuous variable is influenced by a categorical one. Finally, you’ll dive into scatter plots to analyze the relationship between two continuous variables and calculate the correlation coefficient.

In this chapter, you’ll get more familiar with time-based variables and the multiple ways to extract further variables using EDA for analysis—like day of week and time difference. You’ll get hands-on with Power BI as you build line charts to calculate new metrics and uncover trends hiding in your data—including period-over-period change and rolling averages.

One of the most powerful functions of EDA in Power BI is being able to identify which variables have the most influence on your target outcome. To do this you’ll get to grips with two of Power BI’s native visualizations—Decomposition Trees and Key Influencers. First, you'll learn about Decomposition Trees, how to construct, then interpret in order to explain a target outcome by other variables. Finally, you will build a Key Influencers visual, leveraging machine learning, to understand how much a target outcome changes based on specific variables and segments of observations.