Measuring the unmeasurable - a project of domestic violence risk prediction and management
Most common collaborations for domestic violence (DV) prevention in Taiwan are between academic researchers and advocacy groups or governments. However, the limitation of these two types of collaborations often fail to produce effective prevention strategies. In the Data for Social Good project, we performed data-driven research to improve DV prevention strategies using a public-private partnership approach in cooperation with social workers from Taipei City government and a team of voluntary fellows from different professional backgrounds. Fellows were divided into two group to build two prevention tools: a community-based interactive DV risk map and an individual-based repeat victimization prediction model. The risk map visualized complex information in a more comprehensible and accessible manner, facilitating the understanding for its intended users to formulate the specific prevention strategies. On the other hand, the prediction model served as a useful tool for front-line social workers to identify the repeat victimization risk levels of new cases without having to understand complex data. Through this project of public-private collaboration, participants of different roles were motivated by different incentives and rewarded with satisfaction and self-growth that encouraged them to invest even more efforts into the project. This project has attracted a large amount of public attention and initiates a trend of applying data science onto DV prevention among other cities.