Data for Pidgin

Raw Data

The raw data is available here.

The file joint_data.csv contains all the data collected by IfdefRevolver.
The files YYYY-MM-DD/all_functions.csv contains static metrics of all the functions in the respective snapshot.
The files YYYY-MM-DD/ABRes.csv contains static metrics about preprocessor use of all functions in the respective snapshot.
The files YYYY-MM-DD/function_change_hunks.csv contains change metrics of all functions in the respective snapshot.
The files YYYY-MM-DD/snapshot_commits.csv contains hashes of the commits that the respective snapshot encompasses.
The results of the statistical tests for RQ1 are in group_differences.csv
The results of the statistical tests for RQ2 are in spearman.csv
The results of the statistical tests for RQ3 are in logit-reg.csv (logistic regression results), logit-reg-std.csv (logistic regression results after standardization), and nb-reg.csv (negative binomial regression with the outcomes “number of commits” and “lines changes”).

The first picture visuzalizes the distributions of the variables without log transformations.

Distribution of the data without log transformations

The second picture visuzalizes the distributions of the variables after log transformation of the x- and/or y-axis.

Distribution of the data after log transformation of x- and/or y-axes.

The following picures visualize the relationshipt between preprocessor use and functions size.

Relationshipt between number of feature locations (FL) and function size (LOC)

Relationshipt between number of feature constants (FC) and function size (LOC)

Relationshipt between nesting (CND) and function size (LOC)

Relationshipt between negation (NEG) and function size (LOC)

Relationshipt between the proportion of annotated lines of code (LOACratio) and function
size(LOC)