data_analysis_adoption.Rmd

---
title: "Analysis of libraries that declare old dependencies"
output: html_notebook
---

### Import required libraries

```{r}
library(tidyverse)
library(latex2exp)
library(scales)
theme_set(theme_bw())
dim(data_versions)

```

### Read the data

```{r}
data <- read_csv("Data/data.csv")
dataUp <- read_csv("Data/areUpToDate_all.csv")
load("Data/data_versions.RData")
```

### Split the data_versions into four groups

```{r}
# libraries with oly one version released 
libraries_single_version <- data_versions %>% 
  filter(NumVersions == 1)
dim(libraries_single_version)[1]
filtered <- dplyr::anti_join(data_versions, libraries_single_version, by = "Library")

# libraries with mored that one version but released the same day 
libraries_single_day <- data_versions %>% 
  filter(MeanDiff == 0 | SDDiff == 0)
dim(libraries_single_day)[1]
filtered <- dplyr::anti_join(filtered, libraries_single_day, by = "Library")

# libraries that have between 1 and 9 versions released 
libraries_betw10and200_versions <- filtered %>% 
  filter(NumVersions >= 10 & NumVersions <= 200)
dim(libraries_betw10and200_versions)[1]

# libraries that have more than 200 versions released 
libraries_betw10and200_versions <- filtered %>% 
  filter(NumVersions > 200)
dim(libraries_betw10and200_versions)[1]

```


```{r}
table(dataUp$hasNewer)

# group by latest version
data_latest<- dataUp %>% 
  group_by(latest) %>% 
  summarise(Providers=list(provider), hasNewers = list(hasNewer))
dim(data_latest)[1]

# Calculate the number of true
hasNewers <- data_latest$hasNewers

vecTrue <- c()
vecFalse <- c()
for (i in 1:length(hasNewers)) {
  print(i)
  numTrue <- 0
  numFalse <- 0
  for (j in 1:length(hasNewers[[i]])) {
    if (hasNewers[[i]][j] == "true") {
      numTrue <- numTrue + 1
    } else{
      numFalse <- numFalse + 1
    }
  }
  vecFalse <- c(vecFalse, numFalse)
  vecTrue <- c(vecTrue, numTrue)
}

data_latest$NumTrue <- vecTrue
data_latest$NumFalse <- vecFalse

# add colum if the library has at least one outdated dependency or not
data_latest <- data_latest %>% mutate(hasOutdatedDeps = ifelse(NumTrue > 0, "yes", "no"))

# get ratios 
ratioTrue = sum(data_latest$NumTrue) / (sum(data_latest$NumTrue) + sum(data_latest$NumFalse))
ratioFalse = sum(data_latest$NumFalse) / (sum(data_latest$NumTrue) + sum(data_latest$NumFalse))

# data latest
grouped_data <- data_latest
grouped_data$Providers <- as.character(grouped_data$Providers)
grouped_data$hasNewers <- as.character(grouped_data$hasNewers)

write_csv(grouped_data, "data_latest.csv")

# libraries with outdated dependencies 
table(data_latest$hasOutdatedDep)

# libraries with no dependencies
data_latest %>% filter(Providers == "none") %>% nrow()

```

### Reading the data

```{r}
dataa <- read_csv("Data/data.csv")
time_to_release <- read.csv("Data/timeToRelease.csv")
versions <- read_csv("Data/upgrades-new.csv")
metrics_all <- read_csv("Data/artifacts_metrics_all.csv")
latest<- read_csv("Data/latest_versions.csv")
uptd <- read_csv("Data/areUpToDate_all.csv")
```



```{r}
library("ggpubr")

libraries_betw10and200_versions %>% 
  ggplot(aes(MeanDiff, MeanPageranks)) +
  scale_y_log10(breaks = scales::trans_breaks("log10", function(x) 10^x),
   labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  geom_point(alpha = 0.2)

libraries_betw10and200_versions %>% 
  ggplot(aes(MeanDiff, MeanUsages)) +
  scale_y_log10(breaks = scales::trans_breaks("log10", function(x) 10^x),
   labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  geom_point(alpha = 0.2)

```

```{r}
# correlations between number of dependencies and number of versions
libraries_betw10and200_versions %>% 
  ggplot(aes(NumVersions, MeanPageranks)) +
  scale_y_log10(breaks = scales::trans_breaks("log10", function(x) 10^x),
   labels = scales::trans_format("log10", scales::math_format(10^.x))) +
  geom_point(alpha = 0.2)
```