Table 3.
Some of the main concerns generated through a topic modeling approach on the posts mentioning immigrants.
| Theme | Word clouds | ||
|---|---|---|---|
| Self-reference 15.1% |
|
|
|
| Government 12.7% |
|
|
|
| Jobs 11.6% |
|
|
|
| Economy 4.5% |
|
|
|
| Crime and law 6.9% |
|
|
|
| Security 2.3% |
|
||
| Habits and culture 4.02% |
|
|
|
Latent Dirichlet allocation was used to generate 50 topics mentioning immigrants in social media posts. α = 5 favored more topics per post.