Community detection using non-negative matrix factorization

[henripal]

Initially suggested by @hadoopjax in Data4Democracy/discursive#4

• test and implement non-negative matrix factorization using both graph and textual features as described in https://arxiv.org/pdf/1608.01771v1.pdf
• testing to be done on the #far-right twitter dataset

[henripal]

happy to start working on this one!

[henripal]

for the record: we have some results on this but for the next step would be nice to start building a visualization

[ageorgou]

I'd be happy to help with this, whether the visualisation or the algorithmic part.

[hadoopjax]

Hi @ageorgou glad to hear it! @henripal not sure where we left off but would you be able to point @ageorgou in the right direction?

[henripal]

@ageorgou, great! let's talk on stack, I'll point you in the right direction. do you already have access to the data?

[ageorgou]

@henripal I don't, no. I'll message you on Slack!