Book Image

Extending Power BI with Python and R

By : Luca Zavarella
Book Image

Extending Power BI with Python and R

By: Luca Zavarella

Overview of this book

Python and R allow you to extend Power BI capabilities to simplify ingestion and transformation activities, enhance dashboards, and highlight insights. With this book, you'll be able to make your artifacts far more interesting and rich in insights using analytical languages. You'll start by learning how to configure your Power BI environment to use your Python and R scripts. The book then explores data ingestion and data transformation extensions, and advances to focus on data augmentation and data visualization. You'll understand how to import data from external sources and transform them using complex algorithms. The book helps you implement personal data de-identification methods such as pseudonymization, anonymization, and masking in Power BI. You'll be able to call external APIs to enrich your data much more quickly using Python programming and R programming. Later, you'll learn advanced Python and R techniques to perform in-depth analysis and extract valuable information using statistics and machine learning. You'll also understand the main statistical features of datasets by plotting multiple visual graphs in the process of creating a machine learning model. By the end of this book, you’ll be able to enrich your Power BI data models and visualizations using complex algorithms in Python and R.
Table of Contents (22 chapters)
Section 1: Best Practices for Using R and Python in Power BI
Section 2: Data Ingestion and Transformation with R and Python in Power BI
Section 3: Data Enrichment with R and Python in Power BI
Section 3: Data Visualization with R in Power BI

Using AutoML solutions

Writing code from scratch to do machine learning requires specific knowledge that a generic analyst using Power BI often doesn't know. Therefore, we recommend the use of AutoML processes from here on out for analysts who do not have a data science background. Does this mean that anyone can create an accurate machine learning model without knowing the theory behind this science simply by using AutoML algorithms? Absolutely not! The following applies:

Important Note

An AutoML tool relieves the analyst of all those repetitive tasks typical of a machine learning process (hyperparameter tuning, model selection, and so on). Often, those tasks that require specific theoretical knowledge on the part of the analyst (for example, missing value imputation, dataset balancing strategies, feature selection, and feature engineering) are left out of the automated steps. Therefore, not applying the appropriate transformations that only an expert knows to the dataset...