Best Practices
Fire Disaster Prevention
Participating Unit: D4SG (Data for Social Good) Fellowship, Fire Bureau of Kaohsiung City
Sources of Information Used: Fire reporting and fire brigade dispatch records, building permit data, social welfare data, demographic statistics, etc.
Achievements: An AI (artificial intelligence) algorithm was used to build a predictive model of fire risk for buildings, and then households assessed with high risks of fire were identified to significantly improve the success rate of home firefighting visits and the effectiveness of community fire prevention. This is Taiwan's first fire-risk warning model built using AI technology.
Online Dictionary "Moedict"
Participating Unit: g0v "gov zero"
Sources of Information Used: Online version of Mandarin dictionary developed by the Ministry of Education
Achievements: Using the online version of the Mandarin dictionary developed by the Ministry of Education as a foundation to develop a multi-functional, open-source online dictionary for users in social networks and online communities. In recent years, it has covered multiple languages.
Domestic Violence Prevention and Management
Participating Unit: D4SG (Data for Social Good) fellowship, Department of Social Welfare, Taipei City Government
Sources of Information Used: Domestic violence case information, social aids information, demographic statistics, etc.
Achievements: By integrating internal data from the Department of Social Welfare of the Taiwan City Government, a map of domestic violence hotspots has been set up to improve social resources allocation and community advocacy; a relapse risk preventive model for domestic violence has been established to predict the probability of a relapse.
Typhoon Power Outage Prediction
Participating Unit: Taiwan Power Company.
Sources of Information Used: Typhoon power outage reporting data, utility pole information, demographic statistics, meteorological data, etc.
Achievements: Open-source data has been used to organize a data prediction contest. Through crowdsourcing, prediction models have been established, leveraging typhoon power outage data in various regions to assist decision making in disaster prevention and management. The results have been submitted to the power dispatching department for further assessment.