|
Hey there 馃憢 Not sure if this is a silly question but curious to know if regl-scatterplot has the ability to search/highlight scatter labels like in https://projector.tensorflow.org/ - or would that something need to be built? (btw amazing work - repo/examples are awesome!) Zac |
Replies: 2 comments · 1 reply
|
Unfortunately, there isn't any built in functionality for that as Having said that, it should be fairly simple to implement this yourself depending on your data source and structure. For instance, you could index the point labels with something like a TRIE or DAWG (e.g., https://github.com/mckoss/dawg) for fuzzy string matching and then use If you happen to work in with Python, there's a sister application called import jscatter
scatter = jscatter.Scatter(data=df, x="column_a", y="column_b")
scatter.selection(df.query("label == 'my label'").index) |
|
Amazing thanks! |
|
Also, there are no silly questions, so keep them coming! :) |
Unfortunately, there isn't any built in functionality for that as
regl-scatterplotis a bit more "low level" than Tensorflow's Projector app.Having said that, it should be fairly simple to implement this yourself depending on your data source and structure. For instance, you could index the point labels with something like a TRIE or DAWG (e.g., https://github.com/mckoss/dawg) for fuzzy string matching and then use
scatterplot.select(pointIndices).If you happen to work in with Python, there's a sister application called
jupyter-scatter, which integrates into the Pandas ecosystem. That integration makes searching for data points fairly simple.