SPATIAL DYNAMICS OF LOCAL NEWS: MAPPING CITY CO-MENTIONS IN ALABAMA​

This study analyzes 31,004 Alabama news articles (2012–2024) to examine how cities are connected through co-mentions. Using a large language model, we extracted geographic references to build networks capturing spatial and symbolic ties. We developed a relationship typology to interpret these links. Preliminary results show that news categories shape distinct spatial patterns, offering insight into how media narratives influence regional identity and reveal the relational dimensions of news deserts.

1
00:00:07.560 –> 00:00:22.750
Tugce: Hi, everyone! My name is Tugce, I’m a third-year computer science student at George Mason, and my project, Spatial Dynamics of Local News, Mapping City Co-Mentions in Alabama, is about how local news connects cities in Alabama.

2
00:00:22.880 –> 00:00:45.090
Tugce: So, when we talk about news deserts, which is a term in information science that mentions… that describes places that don’t get enough local news coverage, or don’t have enough news outlets to cover their local news, we mostly focus on places that don’t get covered. But something we notice is that news doesn’t just talk about places, it also links them together.

3
00:00:45.340 –> 00:00:57.109
Tugce: For example, two cities might get mentioned in the same story because they share an event, a crime pattern, a rivalry, or a regional issue. These links tell us a lot about how local identity is shaped.

4
00:00:57.280 –> 00:01:01.840
Tugce: And they reveal a lot of information about a region’s new geography.

5
00:01:02.730 –> 00:01:21.400
Tugce: So, what we did is, we took about 31,000 Alabama news articles, ranging from 2012 to 2024, and we used an LLM to plot all the place names, and then built a co-mention network, which is kind of like a map of which cities appear together in stories.

6
00:01:22.950 –> 00:01:30.849
Tugce: We found 79,000 place mentions, 351 Alabama cities, and about 17,000 city pairings.

7
00:01:30.950 –> 00:01:38.430
Tugce: Then we looked at how often each pair of cities co-appears. This gives us an idea of the region’s news geography.

8
00:01:39.040 –> 00:01:51.910
Tugce: So, a few interesting patterns that I like to point out is Tuscaloosa was the main news hub overall, which makes sense because our news outlet is based there. And…

9
00:01:52.140 –> 00:02:09.830
Tugce: We see that Tuscaloosa is the city that gets mentioned together with a lot of other cities in news. And Birmingham becomes the center of crime-related news. So, in crime-related news, the other cities mostly get mentioned together with Birmingham.

10
00:02:10.000 –> 00:02:14.360
Tugce: And Montgomery becomes the center of political and economic coverage.

11
00:02:14.510 –> 00:02:34.510
Tugce: And, not surprisingly, Auburn and Tuscaloosa are super connected in sports stories because of their history of ongoing rivalry. So, what does this all mean? So, when we look at these connections between these cities and how much they appear together in news, we get a story and an idea and a

12
00:02:34.900 –> 00:02:49.240
Tugce: local identity appears about these regions. But when news desserts happen, they aren’t just about where coverage is missing, but also where relationships are missing, because we can’t see these stories.

13
00:02:49.310 –> 00:02:59.169
Tugce: So, some cities don’t just get less coverage, they rarely get linked to other places, which adds another layer of isolation for the news deserts.

19
00:03:32.300 –> 00:03:43.679
Tugce: So, thank you all for listening to my project. I would like to thank my mentors and professors for their invaluable guidance throughout this project. I am especially grateful to my mentor, Dr. Myeong Lee.

20
00:03:43.680 –> 00:03:58.979
Tugce: and Dr. Jieshu Wang for their mentorship, encouragement, and expertise, which shaped the development of this research. I also really appreciate the support of George Mason University’s Oscar program for providing a great research environment.