Detection of traffic events from Finnish social media data
Do Minh, Hang (2016)
2016
MDP in Software Development
Informaatiotieteiden yksikkö - School of Information Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2016-12-07
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:uta-201612152838
https://urn.fi/URN:NBN:fi:uta-201612152838
Tiivistelmä
Social media has gained significant popularity and importance during the past few years and has become an essential part of many people s everyday lives. As social media users write about a broad range of topics, popular social networking sites can serve as a perfect base for various data mining and information extraction applications. One possibility among these could be the real-time detection of unexpected traffic events or anomalies, which could be used to help traffic managers to discover and mitigate problematic spots in a timely manner or to assist passengers with making informed decisions about their travel route.
The purpose of this study is to develop a Finnish traffic information system that relies on social media data. The potential of using social network streams in traffic information extraction has been demonstrated in several big cities, but no study has so far investigated the possible use in smaller communities such as towns in Finland. The complexity of Finnish language also poses further challenges. The aim of the research is to investigate what methods would be the most suitable to analyse and extract information from Finnish social media messages and to incorporate these into the implementation of a practical application.
In order to determine the most effective methods for the purposes of this study, an extensive literature research was performed in the fields of social media mining and textual and linguistic analysis with a special focus on frameworks and methods designed for Finnish language. In addition, a website and a mobile application were developed for data collection, analysis and demonstration.
The implemented traffic event detection system is able to detect and classify incidents from the public Twitter stream. Tests of the analysis methods have determined high accuracy both in terms of textual and cluster analysis. Although certain limitations and possible improvements should be considered in the future, the ready traffic information system has already demonstrated satisfactory performance and lay the foundation for further studies and research.
The purpose of this study is to develop a Finnish traffic information system that relies on social media data. The potential of using social network streams in traffic information extraction has been demonstrated in several big cities, but no study has so far investigated the possible use in smaller communities such as towns in Finland. The complexity of Finnish language also poses further challenges. The aim of the research is to investigate what methods would be the most suitable to analyse and extract information from Finnish social media messages and to incorporate these into the implementation of a practical application.
In order to determine the most effective methods for the purposes of this study, an extensive literature research was performed in the fields of social media mining and textual and linguistic analysis with a special focus on frameworks and methods designed for Finnish language. In addition, a website and a mobile application were developed for data collection, analysis and demonstration.
The implemented traffic event detection system is able to detect and classify incidents from the public Twitter stream. Tests of the analysis methods have determined high accuracy both in terms of textual and cluster analysis. Although certain limitations and possible improvements should be considered in the future, the ready traffic information system has already demonstrated satisfactory performance and lay the foundation for further studies and research.