Building and mining web-based questionnaires and surveys with SySQ.

A questionnaire is a method for collecting data that can come from many sources such as observations, telephone interviews or documentary sources. Whatever the source of data is, the questionnaire provides a framework of questions that facilitate researcher's work. A manual approach for collecting data using questionnaire presents some limitations and introduces several sources of errors. A second issue regards the statistical and data mining of data that often is conducted using different tools than the questionnaire system, which may introduce errors in the analysis pipeline. For instance, common methods applied to data set concern the normality test, the association and correlation discovery, linear regression, classification and clustering. Usually this analysis is performed using external tools, often not free, such as SPSS, SAS, STATA, Weka, or Clementine.We present a web-based software system, to automatize the analysis pipeline and to support researchers involved in the collection of questionnaire data, such as in epidemiology, aiming to reduce the errors listed above and including some basic functions to conduct statistical analysis on collected data. Our system allows researchers to create questionnaires, adding sections and structured questions. It provides a preview of the questionnaire, the exportation of saved data into statistical software compatible formats, or it permits to analyze them directly applying statistical methods and common data mining techniques from the main interface.