Project 2 — Data Tidying: NFL Touchdown Scoring by Type

4. Data Source

This dataset is the official NFL.com team scoring table for the 2026 regular season, pulled from https://www.nfl.com/stats/team-stats/offense/scoring/2026/reg/all. It’s the same kind of dataset Daanish picked for the Week 5 Discussion 5A post: a team by team breakdown of how each team’s touchdowns were scored (rushing vs. receiving vs. two-point conversions), with the touchdown type spread across separate columns instead of stored as a variable.

The raw file, exactly as scraped from the site with no tidying applied, is committed here:

https://raw.githubusercontent.com/nowhyporque/data607DataAcquisitionAndManagement/refs/heads/main/project%202/nfl_scoring_raw.csv

5. Data Structure Before Tidying

nfl_raw <- read_csv("https://raw.githubusercontent.com/nowhyporque/data607DataAcquisitionAndManagement/refs/heads/main/project%202/nfl_scoring_raw.csv")

glimpse(nfl_raw)

Rows: 32
Columns: 5
$ Team   <chr> "Bills", "49ers", "Bears", "Ravens", "Seahawks", "Lions", "Chie…
$ Rsh_TD <dbl> 10, 5, 9, 8, 3, 5, 5, 6, 7, 4, 3, 4, 4, 1, 4, 2, 2, 1, 2, 3, 2,…
$ Rec_TD <dbl> 6, 11, 4, 6, 10, 9, 9, 8, 3, 9, 6, 5, 8, 11, 6, 8, 7, 8, 7, 5, …
$ Tot_TD <dbl> 16, 16, 13, 14, 13, 14, 14, 14, 10, 14, 9, 9, 12, 12, 10, 10, 1…
$ Two_PT <dbl> 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 1, 0, 0, 0, 1, 1, …

kable(head(nfl_raw, 8))

Team	Rsh_TD	Rec_TD	Tot_TD	Two_PT
Bills	10	6	16	0
49ers	5	11	16	0
Bears	9	4	13	0
Ravens	8	6	14	0
Seahawks	3	10	13	0
Lions	5	9	14	0
Chiefs	5	9	14	0
Cowboys	6	8	14	1

This is wide and untidy for the same reason as the example in the assignment: touchdown type is a variable, but instead of living in its own column, it’s spread across the column headers Rsh_TD, Rec_TD, and Two_PT. There are 32 rows (one per team) and 5 columns. Tot_TD is a pre-computed total rather than a raw observation, so it isn’t itself a touchdown type, it gets handled separately in the transformation step below.

6. Transformation Steps

nfl_long <- nfl_raw %>%
  select(-Tot_TD) %>%
  pivot_longer(
    cols = c(Rsh_TD, Rec_TD, Two_PT),
    names_to = "td_type",
    values_to = "count"
  ) %>%

  mutate(
    td_type = case_when(
      td_type == "Rsh_TD" ~ "rushing",
      td_type == "Rec_TD" ~ "receiving",
      td_type == "Two_PT" ~ "two_point_conversion"
    ),
    team = str_to_lower(Team)
  ) %>%
  rename(team_name = Team) %>%
  select(team_name, team, td_type, count)

nfl_long <- nfl_long %>%
  mutate(count = replace_na(count, 0))

kable(head(nfl_long, 9))

team_name	team	td_type	count
Bills	bills	rushing	10
Bills	bills	receiving	6
Bills	bills	two_point_conversion	0
49ers	49ers	rushing	5
49ers	49ers	receiving	11
49ers	49ers	two_point_conversion	0
Bears	bears	rushing	9
Bears	bears	receiving	4
Bears	bears	two_point_conversion	0

7. Analytical Methods

The business question from the discussion post was to compare each team’s scoring style, the ratio of touchdowns by type, not just raw counts. I calculate each team’s rushing and receiving touchdowns as a percentage of total touchdowns (excluding two point conversions, which aren’t full touchdowns), then rank teams by how rush heavy or pass-heavy their scoring is.

scoring_style <- nfl_long %>%
  filter(td_type %in% c("rushing", "receiving")) %>%
  group_by(team_name) %>%
  mutate(total_td = sum(count)) %>%
  ungroup() %>%
  filter(total_td > 0) %>%
  mutate(pct_of_td = round(count / total_td * 100, 1))

rushing_share <- scoring_style %>%
  filter(td_type == "rushing") %>%
  select(team_name, total_td, pct_rushing = pct_of_td) %>%
  arrange(desc(pct_rushing))

kable(head(rushing_share, 5), caption = "Most run-heavy scoring teams")

Most run-heavy scoring teams

team_name	total_td	pct_rushing
Dolphins	4	75.0
Colts	10	70.0
Bears	13	69.2
Bills	16	62.5
Falcons	5	60.0

kable(tail(rushing_share, 5), caption = "Most pass-heavy scoring teams")

Most pass-heavy scoring teams

team_name	total_td	pct_rushing
Giants	7	14.3
Commanders	9	11.1
Raiders	12	8.3
Packers	8	0.0
Steelers	7	0.0

ggplot(scoring_style, aes(x = reorder(team_name, pct_of_td), y = pct_of_td, fill = td_type)) +
  geom_bar(stat = "identity", position = "stack") +
  coord_flip() +
  labs(title = "Share of Touchdowns by Type, All 32 Teams",
       x = "Team", y = "Percent of Touchdowns", fill = "TD Type") +
  theme_minimal(base_size = 8)

The league as a whole leans pass heavy: 62.2% of all touchdowns this season came through the air versus 35.0% on the ground (the remaining share is two point conversions). But individual teams vary a lot. The Dolphins are the most run heavy scoring team, with 75% of their touchdowns coming on the ground, while the Packers are the opposite extreme, 100% of their touchdowns this season have been through the air, with zero rushing touchdowns.

8. Conclusions

Reshaping this table from wide to long made it possible to calculate a real rate (percent of touchdowns by type) instead of just comparing raw counts, which would have been misleading since teams have scored different total numbers of touchdowns. The spread between the most run heavy team (Dolphins, 75%) and the most pass heavy team (Packers, 100% through the air) shows real differences in offensive identity across the league, not just differences in how many points teams have scored overall. A natural next step would be to track this breakdown week by week to see whether a team’s scoring style shifts over the course of the season, for example if a run heavy team becomes more pass reliant once trailing in the standings.