Plotting Fun - Lab Meeting Task

plotting
R
Author

Sam White

Published

November 30, 2023

For today’s lab meeting, Steven tasked us with the following on Wednesday:

For lab meeting Thursday the topic is “Plots”. Before the meeting make sure to create a notebook post where you show-off a plot of something you did in November, including showing the code used to generate the plot. :bar_chart: - Drop the url to the notebook post below.

So, drawing a blank on what to plot, I opted to visualize commit history for the Coral E5 deep dive repo (GitHub repo).

Let’s start with just trying to plot commits per day.

Initial setup

Load libraries


Attaching package: 'dplyr'
The following objects are masked from 'package:stats':

    filter, lag
The following objects are masked from 'package:base':

    intersect, setdiff, setequal, union
Code 
# Check if the '/home/sam/gitrepos/deep-dive' directory exists
if (!file.exists("/home/sam/gitrepos/deep-dive")) {
  # If not, clone the repository
  system("cd /home/sam/gitrepos && git clone git@github.com:urol-e5/deep-dive.git")
}

Set Git Repo Path

repo_path <- "/home/sam/gitrepos/deep-dive/"

Set date range

# Set date range
start_date <- as.Date("2023-11-01")
end_date <- as.Date("2023-11-30")

# Catpures all days in range
all_days <- data.frame(Date = seq(start_date, end_date, by = "day"))

head(all_days)
        Date
1 2023-11-01
2 2023-11-02
3 2023-11-03
4 2023-11-04
5 2023-11-05
6 2023-11-06

Define the Git command to get commit dates using system()

git_command <- paste("git --git-dir=",
                     repo_path, ".git log --pretty=format:%ad --date=short --after=",
                     start_date,
                     " --before=",
                     end_date,
                     " --all",
                     sep="")

print(git_command)
[1] "git --git-dir=/home/sam/gitrepos/deep-dive/.git log --pretty=format:%ad --date=short --after=2023-11-01 --before=2023-11-30 --all"

Data manipulation

Excecute git command

commit_info <- system(git_command, intern = TRUE)

str(commit_info)
 chr [1:233] "2023-11-30" "2023-11-30" "2023-11-30" "2023-11-30" ...

Convert to Dates structure

commit_dates <- as.Date(gsub("'", "", commit_info))

str(commit_dates)
 Date[1:233], format: "2023-11-30" "2023-11-30" "2023-11-30" "2023-11-30" "2023-11-30" ...

Count number of commits

This will create a two-column dataframe (Date, Commits) and replace dates with 0 commits (NAs) with 0.

# This counts number of commits on each date.
commit_table <- table(commit_dates)
str(commit_table)
 'table' int [1:14(1d)] 8 17 19 6 7 21 13 5 4 16 ...
 - attr(*, "dimnames")=List of 1
  ..$ commit_dates: chr [1:14] "2023-11-01" "2023-11-02" "2023-11-03" "2023-11-04" ...
all_days <- merge(
  all_days,
  data.frame(
    Date = as.Date(
      names(commit_table)
      ),
    Commits = as.numeric(
      commit_table)
    ),
  all.x = TRUE,
  by = "Date")

# Assigns value of `0` to all NAs
all_days$Commits[is.na(all_days$Commits)] <- 0

head(all_days)
        Date Commits
1 2023-11-01       8
2 2023-11-02      17
3 2023-11-03      19
4 2023-11-04       6
5 2023-11-05       7
6 2023-11-06      21

Plot all commits per day

ggplot(all_days, aes(x = Date, y = Commits)) +
  geom_bar(stat = "identity", fill = "steelblue") +
  labs(title = "Commits per Day in Nov 2023", x = "Date", y = "Number of Commits") +
  theme_minimal()

Bar plot of all November 2023 commits

Now, let’s try to look at commits per day by author…

Commits per day per author

Create a data frame with all days in November

all_days <- data.frame(Date = seq(start_date, end_date, by = "day"))

New git command

This adds author to format output

git_command <- paste("git --git-dir=",repo_path, ".git log --date=short --all --format='%ad %an'", sep="")

commit_info <- system(git_command, intern = TRUE)

Split the commit information into date and author

commit_info_split <- strsplit(commit_info, " ")
commit_dates <- as.Date(sapply(commit_info_split, "[[", 1))
commit_authors <- sapply(commit_info_split, "[[", 2)

str(commit_dates)
 Date[1:773], format: "2024-09-23" "2024-09-23" "2024-09-23" "2024-09-23" "2024-09-22" ...
str(commit_authors)
 chr [1:773] "Jill" "Jill" "Jill" "Jill" "Jill" "Jill" "Jill" "Jill" "Jill" ...

Create data frame with commits and authors

# Create a data frame with commit information
commit_data <- data.frame(Date = commit_dates, Author = commit_authors)
head(commit_data)
        Date Author
1 2024-09-23   Jill
2 2024-09-23   Jill
3 2024-09-23   Jill
4 2024-09-23   Jill
5 2024-09-22   Jill
6 2024-09-17   Jill
str(commit_data)
'data.frame':   773 obs. of  2 variables:
 $ Date  : Date, format: "2024-09-23" "2024-09-23" ...
 $ Author: chr  "Jill" "Jill" "Jill" "Jill" ...

Update dataframe to include all days in November

all_days <- merge(all_days, commit_data, all.x = TRUE, by = "Date")

str(all_days)
'data.frame':   254 obs. of  2 variables:
 $ Date  : Date, format: "2023-11-01" "2023-11-01" ...
 $ Author: chr  "kubu4" "kubu4" "Steven" "kubu4" ...

Update dataframe with commit counts per author

all_days <- all_days %>%
  group_by(Date, Author) %>%
  summarise(Commits = n())
`summarise()` has grouped output by 'Date'. You can override using the
`.groups` argument.
str(all_days)
gropd_df [45 × 3] (S3: grouped_df/tbl_df/tbl/data.frame)
 $ Date   : Date[1:45], format: "2023-11-01" "2023-11-01" ...
 $ Author : chr [1:45] "Steven" "kubu4" "Steven" "kubu4" ...
 $ Commits: int [1:45] 1 12 4 10 3 5 14 2 4 4 ...
 - attr(*, "groups")= tibble [30 × 2] (S3: tbl_df/tbl/data.frame)
  ..$ Date : Date[1:30], format: "2023-11-01" "2023-11-02" ...
  ..$ .rows: list<int> [1:30] 
  .. ..$ : int [1:2] 1 2
  .. ..$ : int [1:3] 3 4 5
  .. ..$ : int [1:2] 6 7
  .. ..$ : int [1:2] 8 9
  .. ..$ : int [1:2] 10 11
  .. ..$ : int [1:2] 12 13
  .. ..$ : int [1:2] 14 15
  .. ..$ : int 16
  .. ..$ : int 17
  .. ..$ : int 18
  .. ..$ : int 19
  .. ..$ : int 20
  .. ..$ : int 21
  .. ..$ : int [1:2] 22 23
  .. ..$ : int [1:3] 24 25 26
  .. ..$ : int [1:3] 27 28 29
  .. ..$ : int [1:3] 30 31 32
  .. ..$ : int 33
  .. ..$ : int 34
  .. ..$ : int 35
  .. ..$ : int 36
  .. ..$ : int 37
  .. ..$ : int 38
  .. ..$ : int 39
  .. ..$ : int 40
  .. ..$ : int 41
  .. ..$ : int 42
  .. ..$ : int 43
  .. ..$ : int 44
  .. ..$ : int 45
  .. ..@ ptype: int(0) 
  ..- attr(*, ".drop")= logi TRUE

Plot commits per day per author

ggplot(all_days, aes(x = Date, y = Commits, fill = Author)) +
  geom_bar(stat = "identity") +
  labs(title = "Commits by Author per Day in Nov 2023", x = "Date", y = "Number of Commits") +
  theme_minimal() +
  scale_fill_brewer(palette = "Set2")

Stacked bar plot of November 2023 commits per day per author