dplyr ntile() in R: Bin Values into N Equal-Count Groups

The ntile() function in dplyr divides a vector into n approximately equal-count quantile bins, returning an integer 1 to n for each value. It is the rank-based binning function for quartiles, deciles, and any equal-count split.

By Selva Prabhakaran · Published May 12, 2026 · Last updated May 12, 2026

⚡ Quick Answer

ntile(1:10, 4)                      # 4 quartiles
ntile(x, 10)                         # 10 deciles
ntile(desc(x), 4)                    # reverse direction
df |> mutate(quartile = ntile(score, 4))
df |> group_by(g) |> mutate(quartile = ntile(score, 4))
cut(x, breaks = quantile(x, probs = seq(0, 1, 0.25)))  # value-based bins (different)

Need explanation? Read on for examples and pitfalls.

📊 Is ntile() the right tool?

What ntile() does in one sentence

ntile(x, n) returns an integer between 1 and n indicating which quantile bin each value falls into; bins are sized to be approximately equal in COUNT. With 100 rows and n = 4, you get 25 rows per bin.

The standard tool for "split this column into quartiles / deciles / quintiles".

Syntax

ntile(x, n). Returns integers 1..n. Ties are broken arbitrarily by ntile.

Run live

Run live, no install needed. Every R block on this page runs in your browser. Click Run, edit the code, re-run instantly. No setup.

RQuartile bins for a sequence

library(dplyr) ntile(1:12, 4) #> [1] 1 1 1 2 2 2 3 3 3 4 4 4

Three values per bin (12 / 4 = 3).

Tip

ntile(x, n) always returns integers 1 to n. This makes it perfect for grouping or color-coding by quantile. Combine with factor() to get ordered factor labels.

Five common patterns

1. Quartiles (n = 4)

RStandard quartile assignment

ntile(c(10, 20, 30, 40, 50, 60, 70, 80), 4) #> [1] 1 1 2 2 3 3 4 4

Each quartile holds 2 values (8 / 4 = 2).

2. Deciles (n = 10)

R10% bins

ntile(1:100, 10) #> [1] 1 1 1 1 1 1 1 1 1 1 2 2 ...

Useful for "top 10%", "next 10%", etc.

3. Inside mutate as a derived column

RAdd quartile to data frame

mtcars |> mutate(mpg_quartile = ntile(mpg, 4)) |> count(mpg_quartile) #> mpg_quartile n #> 1 1 8 #> 2 2 8 #> 3 3 8 #> 4 4 8

32 rows / 4 = 8 rows per quartile.

4. Reverse direction (highest = bin 1)

RTop performers in bin 1

ntile(desc(c(50, 80, 90, 100)), 4) #> [1] 4 3 2 1

desc() makes the highest value get bin 1.

5. Per-group ntile

RQuartiles within each category

df_g <- data.frame( team = rep(c("A","B"), each = 4), pts = c(10, 20, 30, 40, 50, 60, 70, 80) ) df_g |> group_by(team) |> mutate(quartile = ntile(pts, 4)) #> # A tibble: 8 x 3 #> team pts quartile #> A 10 1 #> A 20 2 #> A 30 3 #> A 40 4 #> B 50 1 #> B 60 2 #> B 70 3 #> B 80 4

Key Insight

ntile(x, n) produces equal-COUNT bins; cut(x, breaks) produces equal-WIDTH bins. They are different. Equal-count is what you want for "top decile" thinking; equal-width is what you want for histogram-like binning. Pick based on whether bin SIZE matters (equal-count) or bin RANGE matters (equal-width).

ntile() vs cut() vs quantile() vs percent_rank()

Four binning / quantile functions in R.

Function	Output	Bin type	Best for
`ntile(x, n)`	Integer 1..n	Equal count	Quartile / decile assignment
`cut(x, breaks)`	Factor	Equal width or custom	Histogram-style binning
`cut(x, quantile(x, probs))`	Factor	Quantile-based	Same as ntile but factor output
`quantile(x, probs)`	Numeric values	(returns thresholds)	Compute percentile boundaries
`percent_rank(x)`	0 to 1	(continuous)	Relative position, not bins

When to use which:

ntile for clean integer bin assignment.
cut if you need factor labels or specific cutpoints.
quantile to find the actual percentile values (e.g., median = quantile(x, 0.5)).
percent_rank for continuous relative position.

A practical workflow

The "quartile-bucketed analysis" pattern is ntile's main use case.

RQuartile-bucketed analysis

df |> mutate(price_q = ntile(price, 4)) |> group_by(price_q) |> summarise( n = n(), avg_revenue = mean(revenue), .groups = "drop" )

Bucket products by price quartile, then aggregate revenue per quartile. Useful for quick "are higher-priced items more profitable" analyses.

For deciles with named labels:

RDecile labels

df |> mutate( decile = ntile(score, 10), decile_lbl = paste0("D", decile) )

Common pitfalls

Pitfall 1: ties broken arbitrarily. ntile uses row order to break ties. Two rows with the same value may end up in different bins. For deterministic ties, sort the data first.

Pitfall 2: NA propagation. NAs in x become NA in the output. Filter NAs before ntile if integer output is required.

Warning

ntile(x, n) does NOT compute true percentiles. Bins are equal-count, not bounded by exact percentile cutpoints. If you need quartile boundaries (Q1 = 25th percentile, etc.), use quantile(x, probs) instead.

Try it yourself

Try it: Bin mtcars$mpg into 5 quintiles (1 = lowest, 5 = highest). Save to ex_quintile.

RYour turn: quintile assignment

ex_quintile <- mtcars |> # your code here count(ex_quintile, mpg_quintile) #> Expected: 5 quintiles, ~6-7 rows each

Click to reveal solution

RSolution

ex_quintile <- mtcars |> mutate(mpg_quintile = ntile(mpg, 5)) count(ex_quintile, mpg_quintile) #> mpg_quintile n #> 1 1 7 #> 2 2 6 #> 3 3 7 #> 4 4 6 #> 5 5 6

Explanation: ntile(mpg, 5) splits the 32 cars into 5 quintiles. 32/5 is not even; some bins get an extra row.

After mastering ntile, look at:

cume_dist() / percent_rank(): continuous relative position
min_rank() / dense_rank(): integer rank
cut(): base R; bins by value cutpoints
quantile(): compute percentile boundaries
forcats::cut_number() / cut_interval(): ggplot-style binning
dplyr::case_when(): custom bin definitions

For ggplot binning, forcats::cut_number() is similar to ntile but returns a factor.

FAQ

What does ntile do in dplyr?

ntile(x, n) divides x into n approximately equal-count bins and returns an integer 1..n for each value indicating its bin. With 100 rows and n = 4, each bin has 25 rows.

What is the difference between ntile and cut in R?

ntile(x, n) produces equal-COUNT bins. cut(x, n) produces equal-WIDTH bins. ntile is for quantile assignment; cut is for value-range binning.

How do I get quartiles with ntile?

ntile(x, 4) returns 1, 2, 3, or 4 per value. The smallest 25% are in bin 1; the largest 25% are in bin 4.

Does ntile handle ties?

Yes, but ties are broken arbitrarily by row order. Two rows with the same value may end up in different bins. Sort first for determinism.

How do I reverse ntile direction?

Wrap in desc(): ntile(desc(x), n). The highest values get bin 1.

Navigate

Tidyverse packages

Deep dives

Wrangling & EDA

Statistics

Machine Learning

Time Series

By Industry

Reporting & Apps

Levels

dplyr ntile() in R: Bin Values into N Equal-Count Groups

What ntile() does in one sentence

Syntax

Five common patterns

1. Quartiles (n = 4)

2. Deciles (n = 10)

3. Inside mutate as a derived column

4. Reverse direction (highest = bin 1)

5. Per-group ntile

ntile() vs cut() vs quantile() vs percent_rank()

A practical workflow

Common pitfalls

Try it yourself

FAQ

Navigate

Tidyverse packages

Deep dives

Wrangling & EDA

Statistics

Machine Learning

Time Series

By Industry

Reporting & Apps

Levels

dplyr ntile() in R: Bin Values into N Equal-Count Groups

What ntile() does in one sentence

Syntax

Five common patterns

1. Quartiles (n = 4)

2. Deciles (n = 10)

3. Inside mutate as a derived column

4. Reverse direction (highest = bin 1)

5. Per-group ntile

ntile() vs cut() vs quantile() vs percent_rank()

A practical workflow

Common pitfalls

Try it yourself

Related dplyr functions

FAQ