https://generativecollective.com/categories/analytics/ Recent content in analytics on Artistic Portfolio of Chris Ried https://generativecollective.com/%3Clink%20or%20path%20of%20image%20for%20opengraph,%20twitter-cards%3E https://generativecollective.com/%3Clink%20or%20path%20of%20image%20for%20opengraph,%20twitter-cards%3E Hugo -- gohugo.io en Sun, 18 Oct 2020 00:00:00 +0000 https://generativecollective.com/posts/using-analytics-to-determine-ideal-hashtags/ Sun, 18 Oct 2020 00:00:00 +0000 https://generativecollective.com/posts/using-analytics-to-determine-ideal-hashtags/ library(stringr) library(readr) library(tidyverse) library(lubridate) I have been curious on what makes an interesting post on instagram based on a larger dataset of images that have been tagged with #generativeart. Some of this is just data discovery, this could seem that there may be a correlation between the tags that have been used and the amount of likes there are. # Extract hashtags patt <- regex("#\\S+") genart <- read_csv("~/InstaCrawlR/table-generativeart-2020-10-11 13:35:33.csv") genart_db <- genart %>% select(ID, Likes, Owner, Date, Text) %>% mutate(hashtags = str_extract_all(Text,patt) ) genart_db_table <- genart_db %>% unnest(cols = "hashtags") %>% mutate(Year = year(Date), Month = month(Date), DayOfWeek = wday(Date), Day = day(Date), Hour = hour(Date)) genart_db_table %>% head() That will generate a rather large dataset