02-03-2017 11:15 AM
I'm trying to predict how often a user is likely to tweet based on how often the user has tweeted in the past. For starters, I want to just keep it super simple–I have a spreadsheet of how often the user has tweeted in the past that includes dates and times, and I would like to model how often the user is likely to tweet in the next few days.
Once I have that figured out, I'd like to add some more complexity to it, like testing whether the sentiment of a particular trending topic influences the likeliness of the user tweeting.
I'm not great at math, but I'm eager to learn how to setup these kinds of models. Can anyone hold my hand a little and get me started?
02-03-2017 11:28 AM
If you just want to look at date/time instances, you should check out some of the tutorial resources on time series and forecasting. The fact that you are looking at tweets is not really relevant if you aren't going to do any text analysis. If you search the forums for these topics, you will find a lot of helpful starters.