I built an app around datacompy #372
Comments
paddymul
added a commit
to paddymul/buckaroo
that referenced
this issue
Jan 16, 2025
# Release 0.8.2 This release makes it easier to build apps on top of buckaroo. Post processing functions can now hide columns CustomizableDataflow (which all widgets extend) gets a new parameter of `init_sd` which is an initial summary_dict. This makes it easier to hard code summary_dict values. More resiliency around styling columns. Previously if calls to `style_column` failed, an error would be thrown and the column would be hidden or an error thrown, now a default obj displayer is used. [Datacompy_app](capitalone/datacompy#372) example built utilizing this new functionality. This app compares dataframes with the [datacompy](https://github.com/capitalone/datacompy) library
|
@paddymul Sorry for the delay. This is great! appreciate you using and building some useful tools around datacompy. I like the idea of having something visual like you've implemented here, just as another QoL for end users. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
This isn't a bug, more a request for collaboration.
I built an app around datacompy leveraging my project - buckaroo.
Datacompy_Buckaroo_2.mov
This comparison tool shows histograms for both dataframes, and highlights values that don't match. If you hover over one of these values, you see the df_2 value for that cell.
In addition if you click below "post processing", you can toggle through different datacompy report sections.
I'd be interested to hear feedback. What type of features would you all want in a quick data comparison tool like this?
source code for the demo
Some caveats:
This currently only supports pandas. Buckaroo does support polars.
Dataframes must be the same length. I'm not doing any type of joining in the main view.
The text was updated successfully, but these errors were encountered: