Custom Histogram Bins Using the breaks Parameter

%useLatestDescriptors
%use dataframe
%use lets-plot(output="js, svg")

LetsPlot.getInfo()

Lets-Plot Kotlin API v.4.12.0. Frontend: Notebook with dynamically loaded JS. Lets-Plot JS v.4.8.1.
Outputs: Web (HTML+JS), Static SVG (hidden)

val url = "https://raw.githubusercontent.com/JetBrains/lets-plot-docs/refs/heads/master/data/diamonds.csv"
val df = DataFrame.readCSV(url)
val data = df.toMap()
println("${df.rowsCount()} x ${df.columnsCount()}")
df.head()

53940 x 10

DataFrame: rowsCount = 5, columnsCount = 10

carat	cut	color	clarity	depth	table	price	x	y	z
0,230000	Ideal	E	SI2	61,500000	55,000000	326	3,950000	3,980000	2,430000
0,210000	Premium	E	SI1	59,800000	61,000000	326	3,890000	3,840000	2,310000
0,230000	Good	E	VS1	56,900000	65,000000	327	4,050000	4,070000	2,310000
0,290000	Premium	I	VS2	62,400000	58,000000	334	4,200000	4,230000	2,630000
0,310000	Good	J	SI2	63,300000	58,000000	335	4,340000	4,350000	2,750000

Default Bins

letsPlot(data) { x = "price" } +
    geomHistogram(color = "black", fill = "gray80")

Equi-probable Bins

val prices = data["price"]
    ?.filterIsInstance<Number>()
    ?.map(Number::toDouble)
    ?.sorted()
    ?: emptyList()
    
val n = prices.size
val priceBins = (0..10).map { i ->
    val pos = i * (n - 1) / 10.0
    val lower = prices[floor(pos).toInt()]
    val upper = prices[ceil(pos).toInt()]
    lower + (upper - lower) * (pos - floor(pos))
}

letsPlot(data) { x = "price" } +
    geomHistogram(
        breaks = priceBins,
        color = "black",
        fill = "gray80"
    ) +
    themeClassic()

Identity Stat

val breaks = (0..15).map { 2.0.pow(it) }
val centers = breaks.zip(breaks.drop(1)) { a, b -> (a + b) / 2.0 }

fun binCenter(v: Double?, breaks: List<Double>): Double? =
    if (v == null || v < breaks.first() || v > breaks.last()) null
    else breaks.zipWithNext().firstOrNull { (a, b) -> v >= a && v < b }?.let { (a, b) -> (a + b) / 2.0 }

val dfWithBin = df
    .convert { "price"<Number?>() }.to<Double?>()
    .add("bin_x") { binCenter(this["price"] as Double?, breaks) }   // bin center, corresponding to current price

val aggDf = dfWithBin
    .groupBy("bin_x")
    .aggregate { count() into "count" }                             // aggregated dataframe: bin center -> size of bin

aggDf

DataFrame: rowsCount = 7, columnsCount = 2

bin_x	count
384,000000	1997
768,000000	13008
3072,000000	10482
6144,000000	11608
12288,000000	6369
24576,000000	979
1536,000000	9497

letsPlot(aggDf.toMap()) {
    x = "bin_x"
    y = "count"
} + geomHistogram(
    stat = Stat.identity,
    breaks = breaks,
    color = "black",
    fill = "gray80"
) + xlab("price")