influencer.Rmd

---
title: "vaccination_influencer"
output:
  pdf_document: default
  html_document: default
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

Find out who are the most influential spokespeople for #vaccine regarding the anti-vax controversy in the US. 
Work in progress - pending the collection of more tweets and further refinements.

```{r}
library(rtweet)
library(igraph)
library(tidyverse)
library(tidytext)
library(hrbrthemes)
library(ggraph)

```

```{r}
####### credentials here
```

## collecting tweets. Needs to be expanded to more tweets and filtered to relevant topics

```{r}
vaccine_tweets<-search_tweets("#vaccine", n=18000, 
                              include_rts=T, geocode = lookup_coords("usa"))
```


# retweet network
```{r}
rt_g <- filter(vaccine_tweets, retweet_count > 0) %>% 
  select(screen_name, mentions_screen_name) %>%
  unnest(mentions_screen_name) %>% 
  filter(!is.na(mentions_screen_name)) %>% 
  graph_from_data_frame() 

summary(rt_g)
```

```{r}
ggplot(data_frame(y=degree_distribution(rt_g), x=1:length(y))) +
  geom_segment(aes(x, y, xend=x, yend=0), color="blue") +
  scale_y_continuous(expand=c(0,0), trans="sqrt") +
  labs(x="Degree", y="Density (sqrt scale)", title="Degree Distribution") 
```

```{r}
i=20 #filter out those too low
V(rt_g)$node_label <- unname(ifelse(degree(rt_g)[V(rt_g)] > i, names(V(rt_g)), "")) 
V(rt_g)$node_size <- unname(ifelse(degree(rt_g)[V(rt_g)] > i, degree(rt_g), 0)) 
```


```{r}
ggraph(rt_g, layout = 'linear', circular = TRUE) + 
  geom_edge_arc(edge_width=0.125, aes(alpha=..index..)) +
  geom_node_label(aes(label=node_label, size=node_size),
                  label.size=0, fill="#ffffff66", segment.colour="green",
                  color="blue", repel=TRUE, fontface="bold") +
  coord_fixed() +
  scale_size_area(trans="sqrt") + 
  theme_minimal()
```


## top hashtags used concurrently with #vaccine
```{r}
vaccine_tweets %>%
  select(hashtags) %>% 
  unnest() %>% 
  mutate(hashtags = tolower(hashtags)) %>% 
  count(hashtags, sort=TRUE) %>% 
  filter(hashtags != "vaccine") %>% 
  top_n(10)
```

## summing followers of followers to gauge primary influence
```{r}
primary_influence <- function(user, trans=c("log10", "identity")) {
  
  user <- user[1]
  trans <- match.arg(tolower(trimws(trans[1])), c("log10", "identity"))
  
  user_info <- lookup_users(user)
  
  user_followers <- get_followers(user_info$user_id)
  uf_details <- lookup_users(user_followers$user_id)
  
  primary_influence <- scales::comma(sum(c(uf_details$followers_count, user_info$followers_count)))
  
  filter(uf_details, followers_count > 0) %>% 
    ggplot(aes(followers_count)) +
    geom_density(aes(y=..count..), 
                 color="lightslategray",
                 fill="lightslategray",
                 alpha=2/3, size=1) +
    scale_x_continuous(expand=c(0,0), trans="log10", labels=scales::comma) +
    scale_y_comma() +
    labs(
      x="Number of Followers of Followers (log scale)", 
      y="Number of Followers") +
    theme_minimal()-> gg
  
  print(gg)
  
  return(invisible(list(user_info=user_info, follower_details=uf_details)))
  
}
```

```{r}
primary_influence(user)
```

```{r}
primary_influence('vaccineepidemic')
```

```{r}
primary_influence('VaccineXchange')
```

### work in progress