This repository extends the statistical analysis capabilities of ReLiS : a tool for conducting systematic reviews, develop by the GEODES software engineering group at the University of Montreal.
It provides to the users a comprehensive Python environment to performs the statistical analysis of their ReLiS SR project classification data.
It uses a Domain Specific Modeling approach (DSM), including a Domain specific language (DSL), to generate statistical analysis models from the SR project classification data. These models are then used to automatically generate the executable artifacts used in the Python environment.
Primary authors: Louis Lalonde @louislalonde and Hanz Schepens @Wickkawizz
- DSL metamodel with ecore
- ReLIS project statistical analysis modelization (from database)
- R statistical functions migration to Python
- Template-based Python artifact generation
| Statistics category | Name | Variable type | Migrated to Python |
|---|---|---|---|
Descriptive |
Frequency tables | Nominal | Yes |
Descriptive |
Bar plots | Nominal | Yes |
Descriptive |
Statistics | Continuous | Yes |
Descriptive |
Box plots | Continuous | Yes |
Descriptive |
Violin plots | Continuous | Yes |
Evolution |
Frequency tables | Nominal | Yes |
Evolution |
Evolution plots | Nominal | Yes |
Comparative |
Frequency tables | Nominal | Yes |
Comparative |
Stacked bar plots | Nominal | Yes |
Comparative |
Grouped bar plots | Nominal | Yes |
Comparative |
Bubble charts | Nominal | Yes |
Comparative |
Chi-squared test | Nominal | Yes |
Comparative |
Shapiro Wilk's correlation test | Continuous | Yes |
Comparative |
Pearson's correlation test | Continuous | Yes |
Comparative |
Spearman's correlation test | Continuous | Yes |