1
$\begingroup$

I wonder if anyone knows of a tool that allows to visualize how a dataset has evolved over time.

There are tools for tracking experiments (neptune, mlflow, etc.) but they are based on keeping track of the parameters that have been used in different runs. I would be interested in tracking whether samples have been added, edited or deleted over time.

Thanks in advance.

$\endgroup$

2 Answers 2

2
$\begingroup$

You could take a look at Data Version Control (DVC), see respective github repo. The functionality partly overlaps with mentioned model/experiment tracking systems.

$\endgroup$
1
$\begingroup$

CometML will show you the data lineage and versioning of the data in the free community version. Won't tell you the diff between versions, but at least I know the order of my versions and which version is the latest.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.