I'm learning how to use graph databases, and asked a question on Stack Overflow about how to represent different views involving some semantically-equivalent objects. I felt this was a valid and thoughtful question, but the reception I got felt very cold and dismissive.

The only (remaining) responses I received insisted that I needed to determine in advance how I would analyze the data, before I could decide how to structure that data. This contradicts my belief in the possibility (and utility) of canonical data formats. I understand that it's possible to optimize data layout in support of some particular analysis (or other usage); however I don't want to tightly-couple my data to some preconceived methodology. I want to represent my dataset with high fidelity to my common-sense, everyday human conception of that data, build up a comprehensive knowledge graph as a durable corpus resource, and then develop different algorithms to analyze it. I want to accumulate my knowledge over time as a cumulative dataset in a machine-programmable format. I don't want to fracture my datasets into disparate representations, because the underlying use case is the same: summarizing different perspectives about the same facts, and evaluating comparative strength of those perspectives based on available evidence. I want to model they way I reason about my beliefs, and those of others, so I can build a tool to help me advocate my perspective more vigorously, using evidence I accumulate and systematically encode over time in a database.

Is this site (datascience.stackexchange.com) a better place to ask such questions? If not, do you know of another Stack Exchange site that is?

related question: Are questions about structuring data for storage in databases/repositories/xxxx systems on topic?

  • 1
    $\begingroup$ Another possibility could be dba.stackexchange. Or maybe some of your explanation here could be a useful edit to your question. One response from one user doesn't have to be the end of story (even though, unfortunately, with a negative score it may get buried now). $\endgroup$
    – oW_
    Aug 22, 2023 at 22:11


You must log in to answer this question.

Browse other questions tagged .