Book Image

Data Engineering with Alteryx

By : Paul Houghton
Book Image

Data Engineering with Alteryx

By: Paul Houghton

Overview of this book

Alteryx is a GUI-based development platform for data analytic applications. Data Engineering with Alteryx will help you leverage Alteryx’s code-free aspects which increase development speed while still enabling you to make the most of the code-based skills you have. This book will teach you the principles of DataOps and how they can be used with the Alteryx software stack. You’ll build data pipelines with Alteryx Designer and incorporate the error handling and data validation needed for reliable datasets. Next, you’ll take the data pipeline from raw data, transform it into a robust dataset, and publish it to Alteryx Server following a continuous integration process. By the end of this Alteryx book, you’ll be able to build systems for validating datasets, monitoring workflow performance, managing access, and promoting the use of your data sources.
Table of Contents (18 chapters)
1
Part 1: Introduction
5
Part 2: Functional Steps in DataOps
11
Part 3: Governance of DataOps

Initial cleansing of datasets

We now have an initial dataset, which we can keep as a raw dataset but doesn't otherwise provide a source that our end users can use to extract value. For example, when we investigate the public consumer price inflation dataset (which we downloaded in the Integrating public data sources with Download tool use section), all the fields are text fields because the reference is a CSV text file. In contrast, the Google Places API data is a complete JSON file but not arranged into usable tables. In both situations, applying any statistical process controls (SPCs) is difficult as the data type doesn't allow for the appropriate statistical measure.

To cleanse our dataset, we will use a generic cleansing process. We will take the concepts needed to cleanse a dataset and apply them to our example raw file.

A simple cleansing process

The preview of our dataset in Figure 4.15 shows four initial problems that we want to address:

  • Titles...