2/2/2023: Daniel O'Brien

[GabeNicholson]

Comment below with a well-developed question or comment about the reading for this week's workshop.

If you would really like to ask your question in person, please place two exclamation points before your question to signal that you really want to ask it.

Please post your question by Tuesday of the coming week, at 11:59 PM. We will also ask you all to upvote questions that you think were particularly good. There may be prizes for top question-askers.

[yjhuang99]

Hi Professor O'Brien, thanks for coming to our workshop and sharing this interesting paper with us! Tobler's First Law of Geography states that everything is related to everything else but near things are more related than distant things. I think your work is actually a vivid example that violence in one geographic location (address) would influence streets and neighborhoods. Though based on the study of Boston, I think this result can also be tested in other places, especially in Chicago. I wonder what policy we can come up with based on the findings.

[jinyz1220]

Thank you for sharing your work! I wonder whether you are also interested in investigating the mechanisms behind the persistence and aggravation of violence in certain streets and addresses (for example, why some streets have higher probability to promote persistence of gun violence), and whether the methodologies used can be applied to the study of the mechanisms. I also think this method can help to map geographic information with smaller scale to gender crime, and eventually remind females to avoid high-risk areas precisely.

[HongzhangXie]

Hi Prof. O'Brien, thank you for sharing your research in crime and geography! I'm not very familiar with crime studies, but I am very interested in the topics about using urban informatics from new technology sources to analyse the effect of the communities and geographies factors in multiple scales. When using these types of data, I can find good results of correlation. But there are many parameters in geographies and communities data and many factors that can affect crime. How to use these types of data to make causal analysis effectively?

[jiehanL]

Hi Prof. O'Brien, thank you for presenting your work! I'm wondering what are the interdependent effects of addresses, streets, and neighborhoods on the persistence and aggravation of crime, and what is a framework for studying and describing these cross-scale interactions?

[shenyc16]

Hi Professor O’Brien,

Thank you for sharing this interesting work with us. I find the paper very inspiring in that the multilevel models are contructed based on neighborhoods, streets and addresses, which contain rich information in social interactions. I think this is an innovative design. My question is, on a scientific ground, how will you detect the potential confounders and conduct sensitivity anaylsis over unobserved covariates since crime is a pretty complex social problem?

[yunshu3112]

Hi Professor O'Brien,

This is an impressive paper and I learned a lot from reading it. I wonder whether machine learning or network analysis can be used in the study of criminology. What do you think does computational power contribute to the study of crimes?

[edelahayeUChicago]

Hi Professor,

I am wondering what the ethical implications are of building a model to guide policy but may have other important variables left out?

Furthermore, I am wondering whether the focus on "neighborhood" would lend itself to being analysed in an agent-based modelling framework??

[Emily-fyeh]

Hi Prof. O’Brien, I am curious about the relationship between microgeographic features and proximity/clustering factors. Is there any particular assumption of crime-prevailing neighborhoods/streets in this particular social context? Also, is there any feature that lead to drastic crime rate changes in two neighboring areas?

[JerryCG]

Hi Prof. O'Brien, When we consider the interaction terms, we have measurements coming from 3 geographical scales, which I think can have significant correlations among them. Will this affect the results or can we extract the independent parts from them? Best, Jerry Cheng

[kuitaiw]

Hi Prof. O'Brien, Thank you for your speaking.And I'm wondering how you rule out other variables when it comes to data at the address level? That is how did you set the control variable? Thank you so much!

[YutaoHeOVO]

Dear Prof. O'Brien,

Thank you for sharing your research with us. Just a few questions that I am not quite sure and I will be grateful if you can further clarify.

The first thing is about the classification on these disorders. We have physical disorders, social disorders and violent crimes. And different indices' predictive power varies across different geographical scales. And I wonder if such heterogeneity in predictive power is the result of the geographical scale itself? (At different levels of geographical scale, the social problem shall be reflected in slightly different way. At parcel level, there might be more domestic problems yet on neighborhood level there can be large-scale public security problem. And I am curious if the difference in results across scales can be interpreted this way.

And another problem that I am curious is that if we accept the idea that the high crime rate is the result of a series of socioeconomic factors, then as long as there is no large-scale treatment to solve these issues, it is unlikely the crime problems will go away themselves. And it seems that such theory will directly lead to an intuitive result that we can observe crime cluster and aggravation in certain urban districts.

And moreover, the selective migration - people who are not satisfied with the crime problems and have the ability to move, yet people who are not able to move will stay, is likely to make the situation even worse.) There are also papers discussing how growing up in a high crime neighborhood affects youth criminal behavior (Damm & Dustmann, 2014) and there are some discussing how migration will affect such criminal behavior.(Chyn 2018) Could you introduce how this research framework accommodate the possible effect of migration which changes the demographic feature and whether such effect will bring any changes in the research?

Thank you so much! Best, Yutao

[hsinkengling]

Hi Professor O'Brien. Thanks for sharing your work with us. I'm wondering if it would make sense to also include crime statistics of neighboring districts as controls, or mediation between the interaction between different scales?

[Peihan12]

Greetings, Dr. O'Brien We appreciate you sharing your research ideas with us. The study appears to explain the potential causes of the formation of criminal conduct in a specific area. I'm wondering whether you're also interested in looking into the factors that contribute to the persistence and escalation of violence in particular neighborhoods and streets.

[sudhamshow]

Thanks for your presentation Dr. O'Brien The results seem quite intriguing. I wondered if the results would extend to geographies that are disproportionately poor or lack trust between their neighbors.

[beilrz]

Hi Professor O'Brien. I was wondering. Do you think the methods and approaches in this paper can be expanded to other geographic regions? Is there any issues you think may hinder the reproducibility of the paper? thanks

[mdvadillo]

Hi Professor O'Brein, I was wondering how you chose which models to run and why discard other machine learning methods known for their predictive power over the chosen model? thank you for your presentation!

[ValAlvernUChic]

Hello Professor O'Brien, thank you for the paper! Like many of my classmates, I wondered how this model might apply to other geographical or urban locations. If we're assuming that Boston has a special set of preconditions, how different of a result can we expect from places like Chicago, LA, New York and perhaps even those outside of the west (Shanghai, singapore, etc. ) and what model assumptions need to be changed for those to work?

[mingxuan-he]

Hello Professor O'Brien, thank you for your awesome work and your presentation. During you time studying communities, I am curious about whether you found any disproportional impact of the presence of public amenities such as green space, parks, and plazas on different groups within the community. If so how should city planning take such impacts into account when designing public amenities?

[xinyi030]

Dear Prof. O'Brien,

Thanks so much for your awesome work and your presentation. I' m interested in practical applications/ implications of the findings of the study and I would appreciate it if you could further clarify. In light of the results from your study, what are some practical applications or implications that can be drawn regarding the persistence and aggravation of violence and disorder in communities? How can the framework of cross-scale interactions be utilized to better understand and address these issues in real-world settings? Are there any specific policy or intervention strategies that you recommend based on the findings of your research?

[ZHE-ZHANG-0213]

Hello, Professor O 'Brien. Thanks so much for your awesome work and your presentation. I wonder if crime statistics from neighbouring areas are also included as a control

[QichangZheng]

Hello, Prof. O'Brein How did you decide which models to test, and why did you choose the chosen model over alternative machine learning techniques that are considered to be more predictive? I appreciate your presentation.