preface

Put yourself in a position where you get a new analytical question from a business stakeholder. If it’s a question you’ve answered umpteen times already, you should focus on streamlining your analytics or utilizing some business intelligence tool to automate the required report generation. But what if you’re hearing the question for the first time? Where do you start?

In the constantly evolving world of data analytics, the emergence of generative AI, and especially its language-oriented incarnation of large language models (LLMs), has brought forth a new era of powerful tools and techniques we can leverage in the field. This should be no surprise, as these models have, in a very short time, found uses in practically every niche of human-computer interaction, including helping aspiring writers come up with ideas for their novels, writing new recipes for vegan dishes, summarizing books into articles, reviewing and polishing up CVs, and many more. The internet is swarming with innovative uses of language models, and new ones are uncovered daily. To the mix are added frameworks and hybrid solutions combining LLMs with omnifarious traditional software methods, such as LangChain, BabyAGI, Langdock, GradientJ, and LlamaIndex, to name a few (check the web for the current list—this one became obsolete before we finished typing this sentence).