I used a machine model that learns with python to identify patterns, but I will refine that, and created my own SQL databases, we can exchange data
You are viewing a single comment's thread from:
I used a machine model that learns with python to identify patterns, but I will refine that, and created my own SQL databases, we can exchange data
From what I have seen the location data can be quite random, so it is not simple. If you look at my latest posts they link to my Github where I have stored user lists. I try to break them down by region. I will be working on Nigeria next and have some plans to automate that. From the metadata I have seen there are just a few cities that get mentioned much.
One issue I have noticed is that there are two metadata columns in the HiveSQL accounts table and location can be either of them. I think it depends on which front end was used. I really need to check both. Actually for mine json_metadata has what I had on St--mit and posting_json_metadata is the current one. I have used the latter in my queries, but I noticed that I missed some people.
My hope is to get the local communities working on these lists as they can pick up cases that scripts would miss, but I would be interested in working together.