data_analysis_evolution.Rmd

---
title: "Evolution of libraries and versions"
output: html_notebook
---

```{r}
# import required packages
library(tidyverse)
library(lubridate)
library(latex2exp)
library(latex2exp)
theme_set(theme_bw())
```


```{r}
# read the data
data <- read_csv("Data/data.csv")
```

## Introduction

This notebook describes the evolution of Maven Central according to the number of new libraries and new versions added each month.

### Evolution analysis

Plot the monthly evolution of the number of new versions released on Maven Central. 

```{r}
# get a data frame with the number of new versions released each month
df_new_versions <- data %>% 
  select(GroupId, ArtifactId, Version, Release) %>% 
  group_by(month = floor_date(Release, unit = "month")) %>%
  summarize(MontlyReleases = n())

# start from 2015 and remove the last month because it is incompleted 
df_new_versions <- df_new_versions[10:171,]

# plot the line graph
df_new_versions %>% ggplot(aes(x = month, y = MontlyReleases)) + 
  geom_line() + 
  scale_x_date(date_breaks = "2 year", date_labels = "%Y") +
  scale_y_log10(breaks = scales::trans_breaks("log10", function(x) 10^x),
   labels = scales::trans_format("log10", scales::math_format(10^.x))) +
   ylab(TeX("number of new versions ($log_{10}$ scale)")) +
   annotation_logticks(sides = "l") +
   xlab(NULL) 
  
```

Plot the evolution of the number of new libraries released on Maven Central.

```{r}
# get a data frame with the number of new libraries released each month
df_new_libraries <- data %>%
  dplyr::select(GroupId, ArtifactId, Release) %>% 
  unite(Library, GroupId, ArtifactId, sep = ":", remove = T) %>% # unite G and A to get the G:A tuple 
  arrange(Release) %>% 
  distinct(Library, .keep_all= TRUE) %>% # keep only the first release for each library
  group_by(month = floor_date(Release, unit = "month")) %>%
  summarize(MontlyReleases = n())

# start from 2015 and remove the last month because it is incompleted 
df_new_libraries <- df_new_libraries[10:171,]

# plot the line graph
df_new_libraries %>% 
  ggplot(aes(x = month, y = MontlyReleases)) + 
  geom_line() +
  scale_x_date(date_breaks = "2 year", date_labels = "%Y") +
  scale_y_log10(breaks = scales::trans_breaks("log10", function(x) 10^x),
   labels = scales::trans_format("log10", scales::math_format(10^.x))) +
   annotation_logticks(sides = "l") +
   ylab(TeX("number of new libraries ($log_{10}$ scale)")) +
   xlab(NULL) 

```

Put both lines in the same plot and prettify.

```{r}
df_new_versions$Type <- "new versions"
df_new_libraries$Type <- "new libraries"
df_bind <- rbind(df_new_versions, df_new_libraries)

df_bind %>% ggplot(aes(x = month, y = MontlyReleases)) +
  # geom_smooth(method='lm',formula=y~x) +
  geom_line(aes(linetype = Type), size = 0.7) +
  # geom_point(aes(shape = Type), alpha = 0.5) +
  scale_x_date(date_breaks = "2 year", date_labels = "%Y") +
  scale_y_log10(breaks = scales::trans_breaks("log10", function(x) 10 ^ x),
    labels = scales::trans_format("log10", scales::math_format(10 ^ .x))
  ) +
  ylab(TeX("number of releases ($log_{10}$ scale)")) +
  annotation_logticks(sides = "l") +
  scale_color_manual(values = c("red", "blue")) +
  scale_shape_manual(values = c(1, 2)) +
  xlab(NULL) +
  theme(legend.title=element_blank(),
        legend.box.background = element_rect(colour = "black"),
        legend.justification=c(1,-0.3), 
        legend.position=c(0.95,0)
        )
```