What Local Environments Drive Opportunities for Social Events? A New Approach Based on Bayesian Modeling in Dallas, Texas, USA

Round 1

Reviewer 1 Report

I found this paper easy to follow. The research topic is suitable for publication in IJGI. My questions/comments/concerns are the following:

1.      I found the paper's title doesn’t reflect its content well. This is a modeling paper to identify/detect potential locations of social events. The POIs and local sociodemographic data were used as input for the model. Still, the focus of analysis and discussion was not so much on quantifying the effects of these factors. Quantifying the effects of the local environment would require efforts relating to identifying the effects (how, both spatially and non-spatially) and measuring the input-output responding changes.  

 2.      The paper lacks a literature review section to support its methodology design, result discussion, and scientific contributions. GIScience methodology is fantastic if it allows us to capture and model spatial-temporal dynamics of the studying phenomenon. But what do we know about these dynamics regarding the nature of social event occurrence for this study? After reading this paper, I urge more background information about social event locations as suggested by the literature. The authors also provide little information about different methods and how they capture spatial dynamics in social event detection.

 Lacking background information also makes it harder for readers to understand the justification behind the place-event network used in this work (Figure 4 b on page 8). Note that there is another Figure 4 on page 9 – which should be Figure 5. This network directly drives the probability results, but there is no explanation for how and why it was designed. The result will be different using different networks.

 3.      The ACS data used for the model was originally at tract level and disaggregated for 400-meter hexagons. This approach creates a spatial structure containing various groups of hexagons with the same sociodemographic variable values. Most hexagons within a group will have the same probability if they don’t contain any POI. How does this impact the model performance and assessments?

4.      There is a mismatch regarding events between Figure 9c and Figure 1a. Figure 1a  shows locations of the available observed events concentrated at the middle of each circle, while Figure 9c shows many more locations detected as correctly predicted. This is misleading. 

Author Response

Please find the response in the attachments. Thanks.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments for author File: Comments.pdf

The English Language was good. But it needs to be more consistent.

Author Response

Please find the response in the attachments. Thanks.

Author Response File: Author Response.docx

Reviewer 3 Report

This study is an interesting and meaningful research endeavor. Using a Bayesian network model based on POI (Point of Interest) data and socioeconomic data, the authors quantify the probabilistic effects of places and people on the presence of social events in Dallas. The research methodology is innovative and the research logic is clear. The handling of various data variables and their representative meanings are well-described, especially with the use of specific examples to illustrate, making it clearer for readers to understand how the entire study was conducted. Here are several suggestions for revision to be considered by the authors when modifying the article:

1. In the introduction section, it is recommended to supplement discussions on the importance and significance of research on social events, and provide literature to support this, highlighting the importance of this study.

2. The authors finally chose 20 variables to construct the Bayesian network model, including 14 types of POIs and 6 sociodemographic variables. However, these variables were not specifically identified. It is suggested that they should be clearly presented.

3. Some specific findings and insights from this study should be discussed in more depth, such as the specific implications of the results, particularly the predictive capabilities of different types of POIs and demographic socioeconomic characteristics on social events. A discussion on the implications for urban public affairs management, urban planning, and construction should be included.

4. Some details need to be corrected. On line 264, it should be Figure 5 instead of Figure 4. On line 136, the number of social event types should be seven rather than eight. 

Author Response

Please find the response in the attachments. Thanks.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Thank you for addressing my comments.