Trying out different kinds of data analysis and visualisation

[jilltxt]

When we have the CSV file with lots of generated stories (see #2) we can start analysing. Is it possible to extract info computationally?

• Can we try some kind of sentiment analysis?
• Word embeddings?
• Simple word clouds showing the most common words for different languages?
• Identify person names and place names and visualise that somehow?

I also very much want to analyse the last sentence, which is always (I think!) about the lessons learned. Any way to computationally analyse that would be great - but we may need to handcode it. Some words are clearly very frequent, e.g. "friendship", "teamwork", "bravery". Maybe with enough generated stories even just a visualisation showing frequency would work?

My first attempts at data visualisation are in https://github.com/MachineVisionUiB/GPT-stories/blob/main/script/kids-stores-lessons.R - just using the first data I got using the chat interface of ChatGPT and handcoding "lessons learned".