A little while ago I developed a python package meant for small data science teams to communicate the rationale and motivation for decisions in developing and modeling data science projects. The package was called KMDS. This package has an upgrade now. You can input your observations in natural language and the package will take care of tagging it appropriately based on a data science project ontology. Conversely, the natural language search is also available, you can query this tool in natural language.
The updated repository with examples is available here:
https://github.com/rajivsam/kmds
Thank you
Rajiv
Rajiv Sambasivan
Posted on
KMDS with New Features
For further actions, you may consider blocking this person and/or reporting abuse
Top comments (2)
You can watch a video that should give the gist of its functional organization and features here: youtube.com/watch?v=b_zmnyOveEI
You can see a step-by-step break down here:
github.com/rajivsam/kmds_migration...
and
an another example here:
github.com/rajivsam/kmds_migration...
Update: KMDS now supports survival analysis as a model type. Added a third case study using an IT service management dataset — modeling time-to-ticket-resolution across support groups (Kaplan-Meier, with still-open tickets handled as censored observations rather than treated as a plain yes/no classification problem). Available here alongside the SBA and Olist examples:
github.com/rajivsam/kmds_migration
Also wrote up how the toolkit is evolving to work alongside coding agents — the idea is that an agent speeds up the mechanical work, but mapping the business problem to the right ML task is still something a human needs to do:
rajivsam.github.io/r2ds-blog/posts...
rajivsam.github.io/r2ds-blog/posts...
Thank you,
Rajiv