Chat-with-data: unique opportunity to engage with data in a conversational way
Architecture, process and example of chat-with-data application
Generative AI and large language models offer a unique opportunity to engage with data in a conversational way. In the story, I will share my experience in developing a “chat-with-data” application. However, let me first start with the answer to the basic question — why chat with data?
Why chat with data
The ever-increasing volume and complexity of data also mean not all data is converted into value. Unlocking value from data requires more tools, techniques, and algorithms. Even though tools and techniques evolve, a technical interface to data always remains a constraint. Our capability to unlock the value in data is as good as the tool that is in front of us. A programming tool like Python will always require someone to write code to analyze the data. A visual tool like Tableau will always require creating joins between different data tables which are available.
Large language models (LLM) have now provided an alternative to code-based and visual tools for us to unlock value from data. Humans can express themselves more easily in natural language as compared to programming language. This means that we can creatively…
