search

alan-turing-institute / netts

Toolbox for creating networks capturing semantic content of speech transcripts.
https://alan-turing-institute.github.io/netts/
GNU General Public License v3.0
8 stars 1 forks source link

First figure in paper for flow of tool #99

Open OscartGiles opened 3 years ago

carobellum commented 3 years ago

This is the current placeholder figure -please let me know about any ideas!

image

carobellum commented 3 years ago

Pipeline_figure_draft_v1

This is the new pipeline figure - Please let me know what you guys think, any and all comments welcome! Would be particularly useful for me to know whether you find the mapping between what the code actually does and what the figure shows intuitive, because you will be some of the few people to have seen the code enough to comment on that.

I will also use this figure for the documentation pages, so please let me know if you have any feedback with that in mind. For example, currently the figure is largely transparent, which means the contrasts would appear a bit different on pages in dark mode. If you have GitHub in dark mode, you will already be able to see that.

Thanks everyone!

@OscartGiles @Iain-S @SarahMorgan

carobellum commented 3 years ago

Pipeline_figure_draft_v2d

Updated the figure, I think it's more intuitive that way - any thoughts?

SarahMorgan commented 3 years ago

Thanks @carobellum ! That looks brilliant- I agree it works well to have the example on the LHS to start. My only minor comments are 1) please can you make the fonts bigger (especially for the paper) and 2) how about a magnifying glass as the icon for 'finding nodes and edges? Then you could use the current 'finding nodes and edges' icon for 'refining nodes and edges' and get rid of the hierarchical graph icon. Otherwise I think it's perfect! :)

SarahMorgan commented 3 years ago

(icons could also be slightly bigger perhaps?)

carobellum commented 3 years ago

Pipeline_figure_draft_v3d

Thank you for the feedback! Hopefully it's better now? Unfortunately the font size of the CoreNLP output is hard to make any bigger since these are screenshots, but I have made them as big as possible with the current formatting.