Predictive analytics for social innovation and public safety

该研究项目将解决使用数据分析来实现社会创新和提高公共安全的问题。原则上，我们将创建从多个部委积累的政府数据中挖掘知识的技术。该提案的工业合作伙伴是 ISM Canada，该公司是加拿大西部创新 IT 业务解决方案的领先市场提供商。 ISM 提供与移动解决方案、云计算、大数据和分析相关的创新服务。该研究团队由里贾纳大学和 ISM 人员组成，将开发收集、存储、挖掘和可视化政府数据的新技术，这些数据通常具有时空特征。我们将把这些技术应用于多个政府数据库，以更深入地了解政府如何改变向公众提供服务的方式，以提高其有效性并降低成本。在我们的研究中，我们将重点关注五个主要问题。我们将设计和评估用于查找和分析用户模式的算法，以识别政府服务成本最高的用户；寻找并分析服务模式，以量化使用单个政府计划对政府的总体成本和收益的成本和收益；在发现特定社会工作者向用户群体推荐的服务以及这些用户获得的利益的模式后，向政府计划的用户推荐服务。此外，我们还将设计和评估新方法来预测和可视化野火的发生和蔓延，这些方法可用于提高公共安全，并确定一种保护公共数据隐私的新方法，以允许从公共数据中挖掘信息，从而保护该数据所描述的人员的隐私。

This research project will address the problem of using data analytics to enable social innovation and increase public safety. Principally, we will create techniques for mining knowledge from government data accumulated by multiple ministries. The industrial partner for this proposal is ISM Canada, a leading market provider of innovative IT business solutions in Western Canada. ISM offers innovative services related to mobile solutions, cloud computing, big data, and analytics. The research team, which consists of U. of Regina and ISM personnel, will develop novel techniques for collecting, storing, mining, and visualizing government data, often with spatio-temporal features. We will apply these techniques to multiple government databases to gain more insight into how the government can change the way services are delivered to the public to increase their effectiveness and reduce their costs. In our research, we will focus on five main problems. We will design and evaluate algorithms for finding and analyzing user patterns to identify the highest-cost users of government services; finding and analyzing service patterns to quantify the costs and benefits of using individual government programs on the overall costs and benefits to the government; and recommending services to users of government programs after finding patterns in both the services that are recommended to groups of users by particular social workers and the benefits obtained by those users. As well, we will design and evaluate new approaches to predicting and visualizing the occurrence and spread of wildfires that can be utilized to increase public safety, and identify a new methodology for preserving the privacy of public data to permit information to be mined from public data in a way that protects the privacy of the people described by that data.