amirnakar/mork.2020

Statistics on App usage data

MorK.2020

Statistics on App usage data

Short Description: This was a small project to test some experiment results on app usage data for a new app

Data:

Questionnaire answers. Either ranking or time difference measurements

• n = 50 for most parameters
• Rank variables: 5 (easy ; font1 ; font2 ; n.menu1 ; n.menu2)
• Continuous variable : 18 (9 for the old app and 9 for the new app)
• Binary/Categorical variables: gender, app use, laguage and others

Statistics in the code:

1. Test normal distribution: Shapiro Wilkes test ; histograms ; qqplots (for multiple parameters)
2. Test mean differencet : Wilcoxon test ; box plots ; mean difference calculations
3. Test correlations : Spearman ; Pearson ; dotplots
4. Permutation MANOVAs : non-parametric MANOVA alternative for 3 varaibles, using permutation tests (Anderson method) to increase robustness

Vizualisations with ggplot:

• Box plots

• QQ Plots

• correlation plots

• Distribution histograms

• Faceted plots for many parameters

The project was completed in October 2020