Hello @nica and all, I am on holiday until the 4th of October, included. So, I would prefer not to stand in any meeting. However, I have prepared some walkthroughs:
- A first look at the Polish corpus
- A first look at the Czech data
- Experimenting with visualizing gender in co-occurrence networks
I would recommend to ask @hugi to repeat the training, it may be a good investment. With it, you can replicate the gist of the analysis in the first two posts above, and extend them at will (for example, explore ego networks, etc.). I write “the gist” because the way Graphryder computes association depth is slightly different from (though highly correlated with) the one in the paper definition that I use when I work with Tulip.
The analysis by gender is bespoke analysis done for POPREBEL, and not supported by Graphryder.
I would also recommend you provide me with requests for visualizations, or data-related questions, and I will take them on at the earliest opportunity. I will probably still generate a draft comparative analysis between the three different corpora, though comparing three is much harder than comparing two. And about that: @Jan, @Richard, @SantosCardonaPR, can I now consider the coding of the German corpus complete?