text data

The five pitfalls of document labeling - and how to avoid them

Whether you call it ‘content analysis’, ‘textual data labeling’, ‘hand-coding’, or ‘tagging’, a lot more researchers and data science teams are starting up annotation projects these days. Many want human judgment labeled onto text to train AI (via supervised machine learning approaches). Others have tried automated text analysis and found it wanting. Now they’re looking for ways to label text that aren’t so hard to interpret and explain.