Browsing by Subject "Data management"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Call me maybe? It's not crazy! Data collection offices are a good partner in data management
    (2015) Sell, Andrew; Hofelich Mohr, Alicia
    For data management professionals, attention is largely focused on the beginning and ends of the research process, as many researchers are worried about meeting federal requirements for data management plans (DMPs) and are looking for ways to share and archive their data. As a University office specializing in survey and experimental data collection, we have seen how the "middle" steps of data collection and analysis can be influenced by, and be an influence on, these upstream and downstream data management processes. In this Pecha Kucha, we will present relevant data management lessons we have learned from designing, developing, and hosting data collection tools. Challenges of anonymity and paying participants, quirks of statistical files produced by data collection tools, and transparency in the research process are among some of the issues we will discuss. As many of these challenges directly impact later sharing and curation of the data collected, we emphasize that data collection offices can be important partners in data management efforts.
  • Thumbnail Image
    Item
    A Traffic Data Management System for Navigation, Collision Detection, and Incident Detection
    (Minnesota Department of Transportation, 1994-06) Shekhar, Shashi; Hancock, Peter A.
    A traffic data management system is an integral part of an IVHS (Intelligent Vehicle Highway System), which obtains information from road sensors, city maps and event schedules, and generates information to drivers, traffic controllers and researchers. We extend the relational database with abstract data types and triggers to model traffic information in a relational database. Abstract data types are needed to efficiently model spatial and temporal information, since they may create inefficiencies in traditional databases. We use monotonic continuous functions to map the object to disk addresses to save disk space and computation time. A model of spatial data is created to efficiently process moving objects. For IVHS databases, we provide schema that have the relevant abstract data types. We also have a large sample of the relations needed to model IVHS data. Several interesting queries are presented to show the power of the model. Triggers are defined, using rule-definition mechanisms to represent incident detection and warning systems. An efficient physical model with the MoBiLe access method is provided.