Spotlights

Alina Ristea

Facebook Twitter Google Print Friendly and PDF

Alina Ristea received her PhD in 2019 from the University of Salzburg, Austria, in Geoinformatics. She joins BARI as a Postdoctoral Research Associate for the 2019-2020 Academic Year.

BARI: Welcome to BARI! We’re excited to have you joining us. Can you tell us a little about your background?

Alina Ristea: Of course! I did my Bachelors at the University of Bucharest, in Romania, in Geography and Cartography. After my Bachelors, I did a two year Masters in GIS (Geographic Information Systems), also at the University of Bucharest. After that, I worked on a research team for two years before starting to pursue my PhD at the University of Salzburg, Austria, in the department of Geoinformatics. My PhD is titled, Integration and Evaluation of Social Media and Crime Prediction Models. I’ve always studied GIS, but for my PhD I started moving towards Criminology and Geoinformatics.

As the title says, my PhD is about integrating the information from social media to better predict crimes in time and space. I had two main objectives; the first one is to find relationships between different features extracted from social media and integrate them in prediction models. Then, after finding relationships and understanding which crime types can have connections with the social media posts, I moved to the second objective, which was the main prediction. I integrated data from Twitter mostly, because it is the most accessible social media at the moment. I did a lot of text analysis, text processing, space and time analysis, and then crime forecasting. So, forecasting in time and space where crime would happen, which crime type, and when.

BARI: Were you primarily using data from Twitter, or were you able to use other social media data as well?

AR: I primarily used the Twitter data. I wanted to use some Facebook data, but it was more difficult to find. Here in Europe, Facebook is a lot more widely used. In the UK you have a lot of tweets, but besides the UK and maybe a little bit from the Netherlands, it’s not as common as cities in the United States, for example. This is also one of the reasons why my research focused on areas where you can find this type of social media, this type of Twitter data. So I focused on the UK, did some analysis in Vancouver, Canada, and then in some parts of the United States, because that’s where the data was available. I was focusing on predicting crime for sporting events, mostly. There was also an opportunity to use different types of sporting events. I looked at football (soccer) in the UK, hockey in Vancouver, and some basketball in the US.

BARI: Did you find crime spikes during sporting events?

AR: Yes. There are some crime types that happen more around the stadium, or around the city center, or around areas with pubs and restaurants where fans tend to gather during home games. So if you do that analysis a little bit after the game or before the game, you find that in the UK fan behavior is quite violent after a football game. I analyzed hockey in Vancouver because it‘s the most common sport there, and again I found some crime types such as theft from a vehicle that would increase during a game period.

According to the literature, if people tend to park their cars in some areas that are not exactly parking places, and there is no security or there is lower lighting, maybe theft from vehicles would be more common because when games happen people tend to come to the stadium by car. This is different in Europe where we tend to use transportation. So we might have a subway that is open all night when there is a game, for example. In some areas you have trams, you have buses that you would use to go to the stadium, which is different from the US or Canada.

BARI: That makes sense. Have you worked with any police departments or law enforcement to start to implement some of findings, or has it been mostly just academic?

AR: It’s mostly academic. I did talk with some police investigators, but unfortunately I didn’t really implement my research. This is something that I would like to do. I would like to strengthen this relationship with stakeholders. I talked with people but at the moment there is not an implementation in police enforcement.

BARI: What’s next for this research? Are you focusing on implementation, or are there other parts of the research that you want to focus on?

AR: This is also one of the reasons why I was interested in the Postdoc position at BARI, because there are some projects that are related to crime, like problem properties and neighborhood characteristics that are attracting crime. I’m also interested in safety perception, and of course the spatial and temporal analysis. Because after all, as I said, I’m a geographer. I have worked a lot with the space and time data. My plan for the future to still analyze crime and analyze characteristics of the environment of urban areas that can be crime attractors or generators, and introduce them in crime prediction models. Not only the models that I did for my PhD, but also create new ones, and analyze this dynamic data and crime prediction models using social media and also mobile phone data, if possible.

BARI: Thinking about the projects that are ongoing at BARI that you’ve talked to [BARI Director] Dan [O’Brien] about, what are you excited about working on?

AR: We discussed several different projects that are now ongoing at BARI. One of them is about problem properties, I’m very excited about this. This project also discusses the idea of broken windows theory. I know there is also some research out there that Dan already pursued with other students on this, and also the use of 311 data and trying to explain the problems in the city, together with the objective crime data from 911 reports. So, this is one of the projects I will focus on, finding patterns in the city that make problematic areas, or exact locations of buildings. How does this affect the entire neighborhood? Because if we have crime just in some locations on the street, it doesn’t mean that the full neighborhood is problematic.

We want to see this risk for problems, like if there is increased crime, are there increased health issues in the same area? I already worked on an urban blight and safety perception project in Louisiana this year. I was there through a Marshall Plan grant for three months at the beginning of the year. We analyzed some specific neighborhoods that have problems with urban blight. It was not exactly in the same area of problem properties that are analyzed one by one, but we picked them by video. We had video cameras in a car, travelled in different neighborhoods, and digitized each property that has signs of urban blight. Then we connected this with crime data.

So this is kind of like problem properties. But as I said, we aggregated the information. This is a connection between the two projects, the one I did in Louisiana and now this problem property project with Dan attracted my attention, but it’s not the only one. I know there’s also some work in segregation and urban mobility. I also have some interest in residential inequality, how it is distributed in space, and how this is changing over time. I would love to be part of this project, and in general in the projects that are, following urban research and the use of social media or other dynamic data in urban research. Analyzing all this user generated digital data is very interesting too, to analyze in space and time.

BARI: It sounds like there’s some great overlap between your research interests and the projects we have going on! We’re glad you’re here, and we look forward to working together. How can people follow your work and get in touch with you?

AR: I’m on Twitter as @alina_ristea, and people can also find me on ResearchGateGoogle Scholar, or LinkedIn.

Published On: November 12, 2019