SDA: Survey Data Analyzer - Shiny Contest Submission

SDA: Survey Data Analyzer

Authors: P.M.S.S. Kumari, P. Wijekoon

Abstract: This web application, SDA, is an online statistical tool which provides a user-friendly interface to do the statistical analysis for the dichotomous data without having a sufficient programming knowledge. Using this App, the researchers can easily identify the primary factors that affect the presence-absence data based on standard statistical models.

Full Description: The web application of SDA: Survey Data Analyzer, can be used as a statistical tool and makes it easy to identify the primary factors that affect the presence-absence data. The user might only deal with it’s user-friendly interface and do the statistical analysis as a menu-driven software package. The relevant Statistical techniques that a researcher may require to analyze the dichotomous data are included here. Therefore, this App is helpful to those who don’t have a sufficient programming knowledge for analyzing survey data. This application mainly focuses on building a model for prediction and finding factors that affect the presence/absence data.

The App SDA has two main tabs, namely “Factor Identification” and “Group Comparison Test”. Using this App, three different types of regression models (Binary Logistic, Random Forest Model and Lasso) can be fitted to identify the primary factors that affect the presence-absence data. These models can be fitted using sub-tabs inside the “Factor Identification” tab. Basic summary statistics of the relevant dataset can also be done by using the “Preliminary Analysis” tab under the “Factor Identification” tab.

Using “Group Comparison Test” tab, an ANOVA test can be performed to identify whether the mean of relevant variable has significant differences among the different groups. Major assumptions related to ANOVA can also be tested, and if the assumptions are violated, the Kruskal-Wallis Rank Sum test can be performed as an alternative method.

Further, you can perform a one sample t-test if normality assumption is not violated or if you have a large sample. Otherwise, you can perform a Wilcoxon Signed Rank test.


Keywords: Survey, Analyzer, Presence/Absence data, Logistic Model, Random Forest Model, Lasso Model, Group Comparisons
Shiny app: http://shamali.shinyapps.io/surveydataanalyzer
Repo: GitHub - ShamaliSujeewa/SDA: SDA makes it easy to analyze presence-absence data, which acts as online statistical software. It provides a user-friendly interface to do the statistical analysis for the dichotomous data without any knowledge of statistical tools.
RStudio Cloud: RStudio Cloud

Thumbnail:
image

Full image:

1 Like