Data Mining in an Electronic Poll (e-Poll) System

This paper introduces the Final Year Project entitled Data Mining in an
Electronic Poll (e-Poll) System, with the problem being how the use of Data Mining
in an Online Poll System can help managers to obtain statistics of customer feedback
or opinions to help achieve their company objectives. The project's objectives are to
conduct study on the use of Data Mining for an e-Poll system and how it affects the
decision of system owners, to design a Data Mart for the Poll that will allow
effortless management of the system's information, and lastly, to produce a working
prototype of the system that allows capturing and retrieval of poll participation
information from store members. The procedures identified to accomplish these tasks
are by going through literature sources to better understand the proper tools and
technique in the development of the system, by observing current online polling
systems and online stores, and by creating a functional prototype of the e-Poll system
to capture poll participation information in order for analysis to be performed unto
them using Data Mining tools. Through the development of this system, the author
finds that Data Mining offers managers to transform their raw data into useful data,
save time in uncovering data trends and analyze vast amounts of data at a time.
Proper design of the Data Mart using Dimensional Modeling or Star Schema allows
optimal query performance and greater understandability without any loss of
information. The e-Poll design allows correlation of the poll participant's feedback to
their personal information, allowing proper analysis and a gateway for managers to
gather potentially important information from all or a sample of their customers,
regardless of geographical boundaries and/or time.
Keywords: Data mimng, e-Poll system, online stores, Data Mart, segmentation,
classification, analysis.