Humbledata NPM

Goal

Humble Data strives to be an in-memory data wrangler with a small and intuitive API. It's useful if you have a small to medium (thousands or tens of thousands) records returned from a database, and you'd like to massage and wrangle and hustle with the data in memory.

Concepts

You use Humble Data to build a Frame object from any data source of tidy data; typically from a CSV file or a database query result. The Frame object can then be manipulated and queried further using aggregate, selection, sorting, and filtering operations. Aggregate functions return one value. All other operations return a new Frame object. This allows operations to be chained.

Humble Data works best with tidy data sets. Tidy data is data that is arranged such that each row represents one sample, and each column represents one variable. In Humble Data, we call a column a field.

Install

npm install humbledata

Usage

Note that the examples below are in TypeScript.

Building the Frame object

The Frame object, once built, is immutable. You build a Frame object with a Builder.

// build by adding one object at a time
import { Builder } from "humbledata"
const frame = new Builder()
  .addRow({ name: 'foo', size: 10 })
  .addRow({ name: 'bar', size: 30 })
  .build()  

// ...or build from a given array of objects
const data = [
    { name: 'alice', age: 20, height: 170 },
    { name: 'bob', age: 30, height: 180 },
    { name: 'charlie', age: 40, height: 175 }    
]
const frame = new Builder(data).build()

Debugging

Frames can be printed to the console with the print() function:

frame.print()
┌─────────┬───────────┬─────┬────────┐
│ (index) │   name    │ age │ height │
├─────────┼───────────┼─────┼────────┤
│    0    │  'alice'  │ 20  │  170   │
│    1    │   'bob'   │ 30  │  180   │
│    2    │ 'charlie' │ 40  │  175   │
└─────────┴───────────┴─────┴────────┘

Aggregate functions

Aggregate functions return a single value calculated from applying an aggregate function to all rows that have a numeric value for the given field.

const sum = f.sum('age') // sum = 90
const max = f.max('height') // max = 180
const min = f.min('height') // min = 170
const avg = f.avg('age') // avg = 30
const median = f.median('height') // median = 175

Counting functions

The count function counts only rows where the given field value is anything else than undefined.

const sparse = [
    { x: 1, y: undefined },
    { x: 2, y: 30 },
    { x: 3, y: 30 }    
]
const frame = new Builder(sparse).build()
const count = frame.count('y') // count = 2 (not 3)

The distinct function returns the number of distinct (unique), non-undefined, values for a given field.

const distinctX = frame.distinct('x') // distinctX = 3
const distinctY = frame.distinct('y') // distinctY = 1

Grouping

The group function combines grouping and aggregation. It groups data by given field, and then it applies an aggregate function to every item in each group. The resulting Frame has one Row per group.

const gameData = [
    { player: 'eva', points: 80 },
    { player: 'eva', points: 10 },
    { player: 'eva', points: 50 },
    { player: 'bob', points: 90 },
    { player: 'joe', points: 20 },
] 
new Builder(gameData)
            .build()
            .print('Player stats')
            .group('player', 'sum', 'points')
            .print('Total points per player')
     
Player stats
┌─────────┬────────┬────────┐
│ (index) │ player │ points │
├─────────┼────────┼────────┤
│    0    │ 'eva'  │   80   │
│    1    │ 'eva'  │   10   │
│    2    │ 'eva'  │   50   │
│    3    │ 'bob'  │   90   │
│    4    │ 'joe'  │   20   │
└─────────┴────────┴────────┘

Total points per player
┌─────────┬────────┬────────────┐
│ (index) │ player │ sum_points │
├─────────┼────────┼────────────┤
│    0    │ 'eva'  │    140     │
│    1    │ 'bob'  │     90     │
│    2    │ 'joe'  │     20     │
└─────────┴────────┴────────────┘

Filtering

The where function is used to filter out rows based on a condition. The where function returns a new Frame object.

f.where('age', '>=', 30).print()
┌─────────┬───────────┬─────┬────────┐
│ (index) │   name    │ age │ height │
├─────────┼───────────┼─────┼────────┤
│    0    │   'bob'   │ 30  │  180   │
│    1    │ 'charlie' │ 40  │  175   │
└─────────┴───────────┴─────┴────────┘

Splitting

The split function splits one Frame into several new Frames, by grouping on a given field.

const f = new Builder().addRows(peopleData).build().print()        
┌─────────┬───────────┬─────┬─────┬────────┐
│ (index) │   name    │ age │ sex │ height │
├─────────┼───────────┼─────┼─────┼────────┤
│    0    │  'alice'  │ 20  │ 'f' │  170   │
│    1    │ 'charlie' │ 40  │ 'm' │  175   │
│    2    │   'per'   │  2  │ 'm' │   95   │
│    3    │  'lise'   │  3  │ 'f' │  125   │
│    4    │ 'august'  │ 48  │ 'm' │  180   │
└─────────┴───────────┴─────┴─────┴────────┘

const res = f.split('sex')
res.map(r => r.print())
┌─────────┬─────────┬─────┬─────┬────────┐
│ (index) │  name   │ age │ sex │ height │
├─────────┼─────────┼─────┼─────┼────────┤
│    0    │ 'alice' │ 20  │ 'f' │  170   │
│    1    │ 'lise'  │  3  │ 'f' │  125   │
└─────────┴─────────┴─────┴─────┴────────┘
┌─────────┬───────────┬─────┬─────┬────────┐
│ (index) │   name    │ age │ sex │ height │
├─────────┼───────────┼─────┼─────┼────────┤
│    0    │ 'charlie' │ 40  │ 'm' │  175   │
│    1    │   'per'   │  2  │ 'm' │   95   │
│    2    │ 'august'  │ 48  │ 'm' │  180   │
└─────────┴───────────┴─────┴─────┴────────┘