Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Crime Prediction and Prevention based on Claims Data Analysis
Increasing level of crime in Switzerland has implications for the society but also for some industries, such as insurance. To address this problem, this project aims at providing an information system which supports individuals in undertaking preventive measures.
Keywords: data mining; machine learning; information retrieval; crowdsourcing; social good
With an increasing number of theft incidents within the last 5 years, especially in the household burglaries segment, Switzerland is becoming one of the leading European destinations for burglars. As an outcome, there is one burglary happening in every eight minutes. This situation has significant implications for individuals and the society leading towards the increased fear of crime. In addition, it is also affecting the insurance industry resulting in higher number of claims, and potentially leading to higher premiums and churn rates. Therefore, providing solutions to individuals that might help them protect themselves against these attacks and increase their safety is of high importance for public services, such as police departments, as well as for the affected industries.
To address the aforementioned issues, this research project aims at designing an information system that predicts and prevents future crimes by means of Big Data analytics. In addition, crowdsourcing is used to obtain richer dataset. The system will support collective actions and motivate individuals to undertake preventive measures by providing personalized advice for crime prevention.
With an increasing number of theft incidents within the last 5 years, especially in the household burglaries segment, Switzerland is becoming one of the leading European destinations for burglars. As an outcome, there is one burglary happening in every eight minutes. This situation has significant implications for individuals and the society leading towards the increased fear of crime. In addition, it is also affecting the insurance industry resulting in higher number of claims, and potentially leading to higher premiums and churn rates. Therefore, providing solutions to individuals that might help them protect themselves against these attacks and increase their safety is of high importance for public services, such as police departments, as well as for the affected industries.
To address the aforementioned issues, this research project aims at designing an information system that predicts and prevents future crimes by means of Big Data analytics. In addition, crowdsourcing is used to obtain richer dataset. The system will support collective actions and motivate individuals to undertake preventive measures by providing personalized advice for crime prevention.
This project aims at providing support to individuals in undertaking preventive measures by providing relevant information regarding the risk of crime through analysis of past events and forecasting the future incidents. In addition, general and personalized tips would provide a direct guidance on the actionable items which could reduce the risk. The risk calculation will be based on historical data obtained from claims’ records from the Swiss insurance company _Die Mobiliar_. In addition, external data sources will be used, such as statistical data (e.g. demographics, weather, urbanization, etc.), police criminal records, etc. Research topics which will be addressed as a part of this project include:
- Spatio-temporal crime prediction
- Information extraction from incident description in a form of prevention tips
- Hotspot visualization
- Crowdsourcing as an approach towards achieving “public good”
- Digitalization of the “neighborhood watch” concept
This project aims at providing support to individuals in undertaking preventive measures by providing relevant information regarding the risk of crime through analysis of past events and forecasting the future incidents. In addition, general and personalized tips would provide a direct guidance on the actionable items which could reduce the risk. The risk calculation will be based on historical data obtained from claims’ records from the Swiss insurance company _Die Mobiliar_. In addition, external data sources will be used, such as statistical data (e.g. demographics, weather, urbanization, etc.), police criminal records, etc. Research topics which will be addressed as a part of this project include:
- Spatio-temporal crime prediction
- Information extraction from incident description in a form of prevention tips
- Hotspot visualization
- Crowdsourcing as an approach towards achieving “public good”
- Digitalization of the “neighborhood watch” concept
Please send your application including CV, transcripts of grades, motivation letter, and (if available) letter(s) of recommendation to:
- Dr. Gundula Heinatz (gundula.heinatz@mobi.ch)
- Dr. Irena Pletikosa Cvijikj (ipletikosa@ethz.ch)
Please send your application including CV, transcripts of grades, motivation letter, and (if available) letter(s) of recommendation to:
- Dr. Gundula Heinatz (gundula.heinatz@mobi.ch)
- Dr. Irena Pletikosa Cvijikj (ipletikosa@ethz.ch)