Hands-On APIs for AI and Data Science

Chapter 1. Creating APIs That Data Scientists Will Love

Application programming interfaces (APIs) are important to data scientists. But how often are data scientists considered by API designers and developers? Data scientists frequently use APIs as data sources for their work. They have some needs that are different than those of software developers or other consumers. If API producers want to make data scientists happy, they will do well to serve these needs.

How Do Data Scientists Use APIs?

The Anaconda State of Data Science Report found that data scientists spend the bulk of their time performing three main activities: preparing or cleansing data (38%); creating reports, presentations, or data visualizations (29%); and selecting, training, and deploying models (27%). This book demonstrates how data scientists use APIs for these tasks.

Preparing or cleansing data often occurs when data scientists perform exploratory data analysis (EDA) on a new dataset by analyzing its contents, formats, and patterns. Other times, this work is part of a scheduled data pipeline, which is a sequence of software tasks that pull from multiple data sources and reformat or remove errors from the data so that it can be used downstream in visualizations, reports, or models. Data engineer is a another job title for people who specialize in these tasks. Chapter 9 will demonstrate using APIs in exploratory data analysis. Chapter 10 will demonstrate using APIs in data pipelines.

Creating reports and visualizations ...