CityVox: public management support system based on automatic analysis of citizens' audio recordings

• Grant number: 23/09706-5
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: December 01, 2023
• End date: July 31, 2024
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computer Systems
• Principal Investigator: Douglas Roberto de Matos Pimentel
• Grantee: Douglas Roberto de Matos Pimentel
• Company: Dom Rock Serviços em Tecnologia S/A
• CNAE: Desenvolvimento e licenciamento de programas de computador customizáveis ; Consultoria em tecnologia da informação ; Tratamento de dados, provedores de serviços de aplicação e serviços de hospedagem na internet
• City: Santana de Parnaíba
• Associated scholarship(s): 23/17385-4 - Development of the CityVox Results Visualization Interface for Public Managers, BP.TT; 23/17474-7 - Development and Processing of Databases for the CityVox System: From Audio Transcription to Graphic interface, BP.TT

Abstract

With the accelerated growth of the urban population worldwide, it is essential to think about ways to make cities more efficient, sustainable, and safe. One possible solution would be to use artificial intelligence to extract relevant information from the contacts citizens make with their city management, providing more informed decisions that consider the real needs of citizens. In Brazil, for example, these contacts are often made by phone, generating thousands of hours of recorded audio annually. These audios contain valuable information for public management, as they form a rich database of citizen demands and opinions. However, city management does not have the tools to leverage the full potential of these audios. This is because most of them (90\% in some cases) are not classified at the time of registration. In addition, the large volume of audios makes it difficult to listen and classify each one after registration. Therefore, we propose CityVox, a system capable of automatically analyzing recorded audio between citizens and their city management and producing information for decision-making. The goal of this project is to demonstrate the technical-scientific feasibility of building CityVox, which will have at its core machine learning and natural language processing algorithms to extract audio information. We will start building the system with a conceptualization of all its components, and end with the first functional version of CityVox, with all its components integrated. This version will be evaluated quantitatively and qualitatively. Considering only cities with more than 100,000 inhabitants and the absence of direct competitors, we estimate that the domestic market for CityVox is worth R$ 710 million per year. Moreover, with CityVox, it is expected that the voice of citizens will have more relevance in the development of public policies and that cities will have a more effective application of their resources. (AU)