Text Clustering Applied to Daily Writing

I'm in the middle of implementing an analytics feature for patrons, including machine learning algorithms like keyword extraction and text clustering to make sense of a vast quantity of texts.

Even if you just wrote a few words every day for a year, you created a non-negligible body of work. Chances are you already have hidden recurring topics. With these features, cowriters can give you insights on how to group smaller unlabelled posts to produce long-form content.

Judging from my personal results, I found out I already have enough material to write three books: a book on digital nomadism (already writing it), one on developing a writing routine, and another on the maker lifestyle. I'm also starting to develop a corpus on green web development.

If I manage to fine-tune this feature, I could probably make a standalone product. You feed it all your writings and it finds actionable patterns between them.

I could also perform clustering at the sentence or paragraph level, but it probably won't output meaningful results. Well, I can only know if I try.

The key to making writing at The Co-Writers a long-term thing is to manage to bridge the micro and the macro. You keep pushing words every day, and some beautiful day you've got yourself a book thanks to the power of technology.

I also found out the only thing I need to glue a group of texts together is a great story, but that's for another post.