Andrew Komaromy

This research aims to explore methods for geographic analysis and visualization of local news based on geographic references included in stories and photos. Using data developed under The Local News Project at Ryerson University's School of Journalism, news items from the Toronto Star were analyzed for the city of Toronto, Ontario as the study area. The distribution of several popular subject groups was analyzed and compared to the distribution of all news items. This research will allow journalism researchers to analyze stigmatization issues, implications on citizen engagement, equitable access, and the role of news media in developing a sense of place.

The literature from a variety of fields of geography was reviewed to determine the most suitable methods for analyzing point-level news data. The data structure from The Local News Project is explained and modifications were made to enable spatial analytical processing. These include event collection, and spatial joining. Several spatial analysis methods were applied to describe spatial distribution, identifying clustering, and enable map comparison. Spatial distributions are described using proportional symbol mapping, standard deviational ellipses, and kernel density maps, while clustering is identified through the use of Ripley's K, hotspot analysis, and local indicators of spatial association. Tercile classification was applied so that densities of the news items' geographic references could be compared using statistical tests. These include the chi-square test of distributions, kappa, and percent agreement.

The results confirm that most Toronto Star news is concentrated in the downtown core. Spatial analysis methods can support researchers in determining details and differences in news coverage that may not be readily seen by examining the data merely as points plotted on a map. In order to provide a rich and relevant dataset for the future, automation processes are strongly recommended. The relationship between readership and news coverage could also be examined for marketing purposes.