The plot below shows stackoverflow tags complementary to [r] that have been ranked in the top five during the period Jan 1 2010 - Jan 1 2016.

ggplot2 reigns supreme, #1 for the last seven years.

statistics has fallen dramatically, perhaps due to R’s increasing generalization?

data.table grew rapidly through 2011, but was displaced by dplyr in 2016. At the same time plyr’s use has dramatically fallen. shiny (suprising to me) out ranks dplyr.

python has always been heavily associated with the [r] tag. It’s complementary use fell dramatically only to rebound in 2014 (likely along with pandas). Perhaps the Python vs. R language wars are nearing their end or are we just getting started?

library(readr)
library(dplyr)
library(ggplot2)
library(ggthemes)

posix <- . %>% as.POSIXct(origin = "1970-01-01", tz = "UTC")

# See http://data.stackexchange.com/stackoverflow/query/560682/complementary-r-tags-in-2016
read_csv("http://data.stackexchange.com/stackoverflow/csv/707591") %>%
  arrange(creation_year) %>%
  filter(creation_year >= posix("2010-01-01")) %>%
  group_by(creation_year) %>%
  mutate(rank = dense_rank(desc(num_tags))) %>%
  group_by(TagName) %>%
  mutate(best_rank = min(rank)) %>%
  mutate(last_rank = last(rank)) %>%
  filter(best_rank <= 5) %>% ungroup %>%
  arrange(last_rank) %>%
  mutate(TagName = factor(TagName, levels = unique(TagName))) %>%
  ggplot(aes(x = creation_year, y = rank,
             colour = factor(TagName),
             group = factor(TagName))) +
  geom_line(size = 1) +
  geom_point(size = 3) +
  scale_y_continuous(trans = 'reverse', breaks = seq_len(10)) +
  scale_color_hc(name = "[tag]\nby last rank") +
  theme_light(base_size = 20) +
  coord_cartesian(ylim = c(1, 30)) +
  ylab("Rank") +
  xlab("Year") +
  ggtitle("[r] Complementary StackOverflow Tags") +
  theme(legend.title = element_text(size = 16),
        legend.text = element_text(size = 16),
        axis.text.y = element_text(size = 15))

For fun, I also went ahead and ran the same query for Python. This data has a slight modification in that it’s from 2012 onward.

pandas meteoric rise is very cool to see. Of course django is top dog.

It looks like python-3.x finally overtook python-2.7 by the end of 2015!

@statwonk