Gathering data from taxi routes through cities and combining it with how people label locations on social media could help us understand how and why crime rates vary, a new study suggests.
Researchers from Penn State University USA looked at how members of social media site FourSquare tagged points of interest across Chicago including restaurants, shops, nightclubs and public transport stations. They were able to analyse the information alongside data on same-area taxi trips over a three month period that included pickup and drop off times and locations, journey time and total fare.
The researchers say that the analysis of crime statistics that encompass population, poverty, disadvantage index and ethnic diversity can provide more accurate estimates of crime rates compared to traditional techniques based on demographic and geographic data.
Jessie Li, assistant professor of information sciences and technology at Penn State explained how the team’s approach likens taxi routes to internet hyperlinks, connecting different communities with each other. "We had this idea that taxis serve as hyperlinks because people are not only influenced by the nearby location, but they are also frequently influenced by the places they go to," she said.
"For example, your home may be a half hour drive from your work; they are not spatially close. But you spend a lot of time there and you end up being influenced by people, such as your colleagues, there."
One surprising discovery is that the data suggests areas with nightclubs tend to experience lower crime rates – at least in Chicago. The explanation may be that it reflects people’s choices to be there, as they want to go to a nightclub that’s safe, not one that’s dangerous.
Presenting the project’s findings at a data-mining conference in San Francisco this week, Li said that big data projects like this can improve understanding of crime by showing how certain areas are used and why people want to be there. This helps planners make better decisions, as well as allowing communities and police to use their resources to tackle crime more efficiently.
Yet although big data can often show correlations between sources of data and effects – such as crime – that are helpful for making predictions, the sources of data are not necessarily causing the effect, Li added.
"What we see here is a correlation between the taxi and points-of-interest data and crime rates. The data show us the correlation, but scientifically, as far as a cause, we don't know."