I have a confession to make: I suck at small talk. I’m good at big talk. Like, existential crisis big. But I want to make people laugh, not cry.<\/p>\n
That’s why I’m working on mastering small talk. My idea is to have a statistically derived list of of frequently asked questions in casual conversations, and witty responses for each one.<\/p>\n
But how do I get this list?<\/p>\n
For this, I need a dataset of real conversations, especially the ones that are about small talk. It should be big enough to show me what kind of questions and topics people usually chat about. I don’t want any artificial or synthetic datasets for this project.<\/p>\n
By the way, do you know if someone has already made something like this? If there is no existing solution, I’ll use the dataset to make my own. But if it already exists, I can skip the hassle.<\/p>\n
COCA, LDC, and BNC, seem to be either paid or restricted. I’ve also seen some related posts on this subreddit.<\/p>\n
https:\/\/www.reddit.com\/r\/datasets\/comments\/u8etiq\/spoken_conversation_datasets_transcripts_needed\/<\/a> https:\/\/www.reddit.com\/r\/datasets\/comments\/mcwldg\/conversational_datasets\/<\/a> https:\/\/www.reddit.com\/r\/datasets\/comments\/6bjzgl\/i_put_together_a_few_conversational_datasets_if\/<\/a><\/p>\n<\/div>\n submitted by \/u\/8ta4 <\/a>
[link]<\/a><\/span> [comments]<\/a><\/span><\/p>
