Types of data ML/AI devs work with
Understanding the types of data and their applications in machine learning and artificial intelligence
About this Report
The aim of this report is to explore what types of data ML/AI developers work with. To begin, we broadly examine popular types of data and compare our findings between those involved in deep learning technologies and those who aren’t. In chapter two, we examine what influences their choices, while chapter three deep dives into whether data type and size influence the choice of architectures ML/AI developers use to train and deploy models. In the final chapter, we turn our attention to what ML/AI developers use each of these types of data for. Going further, we examine developers working on applications of the two most popular types of data — text and image data — and consider who their primary target audience is and what their main aims for involvement in the sector are.
Key Questions Answered
What types of data do ML/AI developers use?
What is the typical size of the training datasets ML/AI developers work with?
Do the types of data ML/AI developers work with differ by experience in ML/AI development?
Do the size of the training datasets ML/AI developers work with differ by organisation size?
Are ML/AI developers’ preferences for computing architectures to train and deploy models influenced by the type and size of the data that they work with?
What are the most popular ML/AI applications of text and image data?
What are ML/AI developers working on different applications trying to achieve?
Click to expand
Methodology
The data for the report comes from the 24th edition of our Developer Nation survey, which ran between December 2022 and February 2023 and reached more than 6,100 ML/AI developers.