general improvement ideas
The following is a lose list of ideas that might improve the framework and the workflows around it:
-
separate framework from analysis - have one repository where the framework is developed
- create a repository every year which contains data, analysis scripts, results
-
improve usability for windows users (running and developing), includes documentation for windows +1 (cf. #200) -
create an output method for textual data instead of using print statements (cf. #201) -
keep text and pictures closely together via naming convention or auto naming (cf. #202) - it is somewhat hard to identify what the print are showing from the name of the file
-
create proper dev and analysis cycles (which do not overlap) for the framework (cf. #203) -
add general pre processing, which happens before any analysis is started (cf. #204) - sort out empty / bad entries, generate descriptive statistics, ...
- by argument, config file, (hard coded), ...
- document what was done (replacing values, dropping entries, ...)
-
default percentage scales to 0-100 (cf. #205) -
generate short question / answer text (either in the survey itself or the meta data) -
better column names (from short texts?!) (cf. #206, #207) -
HIFIS/Helmholtz corporate design for plots -
organize the three survey projects in one subgroup - ...
feel free to add entries, please discuss every topic in a separate thread or issue
using cc with specific names, because group members are 63 - feel free to add people if i forgot someone
/cc @bezaye @mdolling-gfz @erxleb87 @frust45 @hueser93 @KatDwo @Normo @schlauch