https://github.com/favstats/allsider
Media Bias Ratings from AllSides.com
https://github.com/favstats/allsider
Last synced: 8 months ago
JSON representation
Media Bias Ratings from AllSides.com
- Host: GitHub
- URL: https://github.com/favstats/allsider
- Owner: favstats
- License: gpl-3.0
- Created: 2019-10-17T14:51:06.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2019-10-18T00:02:47.000Z (over 6 years ago)
- Last Synced: 2025-08-13T09:45:14.973Z (10 months ago)
- Language: R
- Size: 231 KB
- Stars: 24
- Watchers: 1
- Forks: 4
- Open Issues: 0
-
Metadata Files:
- Readme: README.Rmd
- Contributing: .github/CONTRIBUTING.md
- License: LICENSE.md
- Code of conduct: .github/CODE_OF_CONDUCT.md
- Support: .github/SUPPORT.md
Awesome Lists containing this project
README
---
output: github_document
---
Latest Data Update: 2019-10-18
```{r, include = FALSE}
knitr::opts_chunk$set(
collapse = TRUE,
comment = "#>",
message = F,
warning = F,
fig.width = 8,
fig.height = 8
)
library(tidyverse)
```
# AllSideR
The goal of AllSideR is to provide you with the [AllSides.com](https://www.allsides.com/media-bias/media-bias-ratings) Media Bias rating dataset.
From AllSides.com:
> Our Media Bias Ratings are determined using multiple methods and represent the average judgment of Americans. They are based on blind surveys of people across the political spectrum, multi-partisan analysis, editorial reviews, third party data, and tens of thousands of user feedback ratings. Our scientifically-generated ratings are fluid and subject to change as new information is gathered and biases change over time.
The dataset looks like this:
+ `news_source`: Rated entity
+ `rating`: AllSides Rating (left, left-center, center, right-center, right, allsides)
+ `rating_num`: Same as `rating` but numeric (from left to right, excluding "allsides" rating)
+ `type`: News entity type (can be news media, author or think tank/policy group)
+ `agree`: Number of people who agree with the rating
+ `disagree`: Number of people who disagree with the rating
+ `perc_agree`: Percentage of people who agree with the rating
+ `url`: URL with more Info
+ `editorial_review`: Was there an editorial review of the rating? (0/1)
+ `blind_survey`: Blind survey conducted? (0/1)
+ `third_party_analysis`: Based on third-party analysis? (0/1)
+ `independent_research`: Based on ndependent research? (0/1)
+ `confidence_level`: The confidence of the rating (Low or Initial Rating , Medium, High)
+ `twitter`: URL to Twitter account (when provided)
+ `wiki`: URL to Wikipedia article (when provided)
+ `facebook`: URL to Facebook page (when provided)
+ `screen_name`: Twitter screen name
## Get Dataset
Either Download data from GitHub Repo like this:
```{r}
allsides_data <- readr::read_csv("https://raw.githubusercontent.com/favstats/AllSideR/master/data/allsides_data.csv")
allsides_data
```
Or install package and use it from there:
```{r, eval=F}
devtools::install_github("favstats/AllSideR")
## Access dataset from package
AllSideR::allsides_data
```
## Example Plots
How does rating relate to agreement of the rating?
```{r}
allsides_data %>%
ggplot(aes(rating_num, perc_agree*100)) +
geom_jitter() +
geom_smooth() +
labs(x = "AllSides Media Bias Rating (Left to Right)", y = "% Agree",
title = "AllSides Media Bias Rating X Agreement with Rating",
caption = "Source: AllSides.com") +
ggrepel::geom_text_repel(data = allsides_data %>% filter(perc_agree*100 > 80, rating_num %in% c(1:5)),
aes(label = news_source)) +
theme_minimal()
```
People agree more on the bias of partisan media.
```{r}
allsides_data %>%
mutate(partisan = case_when(
rating_num %in% c(2, 4) ~ "Slightly Partisan",
rating_num %in% c(1, 5) ~ "Partisan",
rating_num == 3 ~ "Center"
)) %>%
drop_na(partisan) %>%
mutate(partisan = fct_relevel(partisan, c("Center", "Slightly Partisan", "Partisan"))) %>%
ggplot(aes(partisan, perc_agree*100)) +
geom_jitter(alpha = 0.4) +
geom_violin(alpha = 0.4) +
geom_boxplot(width = 0.2) +
labs(x = "AllSides Media Bias Rating", y = "% Agree",
title = "AllSides Media Bias Rating X Agreement with Rating",
caption = "Source: AllSides.com") +
theme_minimal()
```
Finally, let's take a look at the distribution of bias by source `type`.
```{r}
allsides_data %>%
ggplot(aes(rating_num, fill = type)) +
geom_density(alpha = .3) +
theme_minimal() +
theme(legend.position = "bottom") +
labs(x = "AllSides Media Bias Rating (Left to Right)", y = "Density",
title = "AllSides Media Bias Rating",
caption = "Source: AllSides.com")
```
Lack of right-wing media is filled by "Author" category. Would be worth investigating.