Spatiotemporal Analysis of Nighttime Crimes in Vienna, Austria
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsI really enjoyed this paper and it has clear added value to the current literature.
1. Does your crime data include a mix of public place and indoor crimes. I presume so with burglary and assault. I am not sure impact of street lighting is straightforward in terms of increasing/reducing crime as this might vary by crime type. I rink you need to think theoretically about impact of lighting on crime theoretically for your different crime types. See further points below re guardianship. Burglary occurs indoor. Assault is outdoors. Theft I presume is outdoors (is this personal or vehicle for example?)
2. I have some reservations about using average time for identifying crime times, especially for burglary. If no one is home between 5pm and 11pm I think there are problems with using a midway point that might indicate night-time. It would be more appropriate to consider using an aoristic method for estimating the times. See work of Jerry Ratcliffe.. At least for burglary you should use Aoristic analysis.
3. I have concerns about using OLS methods – most crime follows Poisson distributions and usually negative binomial regression is more appropriate. I think use of GRW is good however but I would replace LS with NBR (once tested fit). It may be zero inflated Poisson.
4. How might lighting intersect with capable guardianship – especially when visibility is concerned. There is an interesting balance between better lighting to identify opportunities and better lighting to increase risk of detection/apprehension. Again crime type is important as is local context - see point 5.
5. In residential areas there will few businesses still open during night-time – therefore place management/guardianship will be reduced in darkness (see recent work by Eck and Linning after Jane Jacobs). However, in the night-time economy business will be open during hours of darkness and therefore there will still be guardians present to deter crimes. Transport stations may be open later than playgrounds for example. You should consider this in your limitations and discussion. Some land use types could be included as independent variables
6. Lighting is one factor to have shown to influence crime – but weather does too. See work on robbery by Lisa Thompson and Kate Bowers (Stab in the dark) which also explores rain for example as well as daylight hours. Again this might influence your results and or is a limitation here.
7. There are some recent studies in the UK about impact of turning off street lights on crime (local authorities saving money in times of austerity). There were different impacts on eg theft from vehicle. This might also help with your consideration of changes since 2017.
8. You need to be careful with your definitions. For example night-time lights and street lighting. Please define. I think of street lights as those put in by local authorities and that automatically turn on. There are other lights in places (eg from businesses or houses or cars). At times you mention satellites/remote sensing as way of capturing night-time light - but also a dataset provided on Street lights so I found this confusing. Which have you captured.
9. I think given you have examined so many crime types your findings have become a little confused. Also is the contextual data you use appropriate for each crime type (eg population, race etc). I suspect these variables should be more closely considered for each of your crime types.
10. Take care with writing. There are some sentences that need checking. For example –“First, crime type is one of the important factors that can be used to increase or decrease night-time crimes by NTLs”. How does crime type increase or reduce night-time crimes. I think you mean night time crime varies by crime type but take care with phrasing.)
Comments on the Quality of English Language
Some minor changes are needed. See eg point 10.
Author Response
We appreciate your valuable feedback. Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThe manuscript presents a study on inverstigating the relationship between nighttime light and criminal activities in Vienna, Austria. The idea is really interesting. However, it also has a lot of flaws:
- P1, L34. This sentence is not proper. There have been a lot of studies on nighttime crimes. The following pragraph aslo summaries some of them.
- Section 2. Some of the reviewed studies did not cite references in this section.,
- P6, L215-217. This sentence should be removed becasuse Luojia 1-01 was not employed in this study.
- Fig. 4 The NTL values are extremely high is some months. Please explain this phenomenon.
- Fig. 7. Is not necessary. It's better to remove this figure.
- P14, Equation 4-11 The regresssion equations are not necessary. The authors shall compared the performances of different methods and determine the optimal model with the highest accuracy. Then the analysis is carried out based on the optimal method.
- The result section should be refined.
- A discussion section is required to state the advantages and limitations of this study.
Author Response
We appreciate your valuable feedback. Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe manuscript entitled "Spatiotemporal Analysis of Nighttime Crimes in Vienna, Austria" aimed to investigate the relationship between nighttime crime types utilizing nighttime light data, streetlight and socioeconomic data in Vienna, Austria.
The manuscript is interesting, organized and comprehensive. In my opinion it is suitable to publish in "International Journal of Geo-Information".
I have some comments.
1) Page 6, line 212. The radiometric resolution is not 8 bits.
2) Page 6, line 218. Determine the product type of the VIIRS NTL data.
3) Page 8, line 268. I believe you mean Figure 6 not 5.
4) Page 8, line 269. You mentioned that (…., the raster images are converted to a grid cell to conduct …. ).
What do you mean by “a grid cell”? I think you mean “a grid polygon”. If you mean a grid polygon, change it in the whole manuscript.
5) Page 9, lines 274-275. Put a reference at the end of the sentence.
6) Page 14, line 409. You mentioned that “In general, the median of NTL values shows a higher correlation with nighttime crimes than the streetlight densities”. Based on your results, the reverse is true.
7) Page 14. Delete all equations from 4 to 11. These equations easily can be read them from Table 6. Also delete equations 12, 13, 14 and 15.
8) Page 25, line 631. Discuss the limitation of your method.
Author Response
We appreciate your valuable feedback. Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsThe work reported in the manuscript was well done and presented. It might be useful for the authors to comment on the robustness of the relatively small adjusted R^2 values. Other than that, the work is well done.
Author Response
Thank you for your valuable feedback. We added text regarding the robustness of adjusted R-squared values (please see line 715).
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsThe authors have incorporated my suggested changes which has improved the quality of the manuscript. I have only one remaining comment:
A separate discussion section is still needed. This section should not only discuss the limitations, but also compare this study with previous studies and address the academic contributions of this paper.
Author Response
Added a separate Section “5. Discussions" (lines 682-712) 1) to compare this study with previous studies, 2) to address the academic contributions of this paper, and 3) to discuss the limitation of this study.
