DataFrames & Series

About

Originally inspired by my frustration in the lack of types provided in pandas, this repo is an implementation that has a focus on readability and easy maintenance.

It's JavaScript, so if you're looking for amazing performance you should be looking else where, but if you just want to dip your feet into DataFrames then this is a good place to start.

Series

import { Series } from "../src/series";

const series = new Series([1, 2, 3, 4, 5], "myNumbers");
const mean = series.mean();
/** 3 */
console.log("mean: ", mean);

const median = series.median();
/** 3 */
console.log("median: ", median);

const min = series.min();
/** 1 */
console.log("min: ", min);

const max = series.max();
/** 5 */
console.log("max: ", max);

const sum = series.sum();
/** 15 */
console.log("sum: ", sum);

const concatSeries = series.concat([6, 7, 8, 9, 10]);
/** [1, 2, 3, 4, 5, 6, 7, 8, 9, 10] */
console.log(concatSeries.items);

const stringSeries = concatSeries.lambda((num) => num.toString(), "myStrings");
/** ['1', '2', '3', '4', '5', '6', '7', '8', '9', '10'] */
console.log(stringSeries.items);

const nullishFilledSeries = new Series(
  [1, undefined, false, null, Number.NaN, 10, ""],
  "myNullishMix",
).fillNullish(0);
/** [1, 0, false, 0, 0, 10, ''] */
console.log(nullishFilledSeries.items);

const falseyFilledSeries = new Series(
  [1, undefined, false, null, Number.NaN, 10, ""],
  "myFalseyMix",
).fillFalsey(0);
/** [1, 0, 0, 0, 0, 10, 0] */
console.log(falseyFilledSeries.items);

DataFrame

import { DataFrame } from "../src/dataframe";

const df = new DataFrame([
  { name: "Alice", age: 30, active: true },
  { name: "Bob", age: 25, active: false },
]);

/** [2, 3] */
console.log(df.shape);

/** ['name', 'age', 'active'] */
console.log(df.columns);

const meanAge = df.col("age").mean();
/** 27.5 */
console.log(meanAge);

df.pushRow({
  name: "John",
  age: 35,
  active: true,
});

/** [ { name: 'Alice', age: 30, active: true } ] */
console.log(df.head(1));

/** [ { name: 'John', age: 35, active: true } ] */
console.log(df.tail(1));

const withInitials = df.addColumn("initials", (cols) => cols.name[0]);
/**
 * [
 *  { name: 'Alice', age: 30, active: true, initials: 'A' },
 *  { name: 'Bob', age: 25, active: false, initials: 'B' } ]
 * ]
 **/
console.log(withInitials.head(2));

const dfLeft = new DataFrame([
  { id: 1, name: "Alice" },
  { id: 2, name: "Bob" },
]);

const dfRight = new DataFrame([
  { userId: 1, age: 25 },
  { userId: 3, age: 30 },
]);

const joined = dfLeft.leftJoin(dfRight, {
  thisKey: "id",
  otherKey: "userId",
});

/**
 * [
 *   { id: 1, name: 'Alice', age: 25 },
 *   { id: 2, name: 'Bob', age: null }
 * ]
 */

console.log(joined.head(2));

const assigned = df.assign({
  birthYear: (row) => 2025 - row.age,
  nameLength: (row) => row.name.length,
});

/**[
 *   {
 *      name: "Alice",
 *      age: 30,
 *      active: true,
 *      birthYear: 1995,
 *      nameLength: 5,
 *    },
 *    { name: "Bob", age: 25, active: false, birthYear: 2000, nameLength: 3 },
 *  ]
 */
console.log(assigned.head(2));