See the Video Series Here
This library allows you to work with data structured as a table or as a matrix and provides you with many of the methods you would expect when working with such things. It also provides various convenience and statistical functions.
A dataset represents a collection of object-rows. Among other capacities, here you have the
ability to map, filter, sort, group, reduce, and join. These methods can seem similar to
those found on Array. However, they are designed to work with objects as rows. Furthermore,
some SQL-like capacity (e.g. left join, exists) and deeper statistics (e.g. multiple regression)
are available that you just cannot get in vanilla javacript.
A matrix is a rectangular collection of numbers on which particular mathematical operations
are defined. This library offers many of the expected operations of matrix algebra. This
includes matrix multiplication, addition, various methods of 'apply' functionality, varous
decompositions, pseudoinvering, and production of eigen values and vectors.
Click on the links below to see more information in each area:
- fluent-data the default export from the library
- dataset methods and examples
- matrix methods and examples
- changelog
npm install fluent-data
// client
import $$ from './node_modules/fluent-data/dist/fluent-data.client.js';
// server
let $$ = require('fluent-data');
// but the examples in this documentation will use
let $$ = require('./dist/fluent-data.server.js');
Consider the following arrays:
let customers = [
{ id: 1, name: 'Alice' },
{ id: 2, name: 'Benny' }
];
let purchases = [
{ customer: 2, speed: 15, rating: 50, storeId: 1 },
{ customer: 1, speed: 5, rating: 90, storeId: 1 },
{ customer: 1, speed: 7, rating: 55, storeId: 1 },
{ customer: 2, speed: 6, rating: 88, storeId: 1 },
{ customer: 1, speed: 25, rating: 35, storeId: 1 },
{ customer: 1, speed: 40, rating: 2, storeId: 3, closed: true },
{ customer: 2, speed: 4, rating: 88, storeId: 1 },
{ customer: 1, speed: 1, rating: 96, storeId: 2 },
{ customer: 1, speed: 2, rating: 94, storeId: 2 },
{ customer: 1, speed: 1, rating: 94, storeId: 2 }
];
The following example converts the to dataset and uses many of the methods available.
let $$ = require('./dist/fluent-data.server.js');
$$(purchases)
.filter(p => !p.closed)
.joinLeft(customers, (p,c) => p.customer == c.id)
.group(p => [p.customer, p.storeId])
.reduce({
customer: $$.first(p => p.name),
store: $$.first(p => p.storeId),
orders: $$.count(p => p.id),
speed: $$.avg(p => p.speed),
rating: $$.avg(p => p.rating),
correlation: $$.cor(p => [p.speed, p.rating])
// other reducers, such as multiple regression, are built in!
})
.sort(p => [p.customer, -p.rating])
.log(null, 'purchases:',
p => $$.round({ ...p, orders: undefined}, 1e-3)
);
// use 'get' as opposed to 'log' to assign to a variable
This results in three rows for analysis:
purchases:
┌──────────┬───────┬────────┬────────┬─────────────┐
│ customer │ store │ speed │ rating │ correlation │
├──────────┼───────┼────────┼────────┼─────────────┤
│ Alice │ 2 │ 1.333 │ 94.667 │ -0.5 │
│ Alice │ 1 │ 12.333 │ 60 │ -0.832 │
│ Benny │ 1 │ 8.333 │ 75.333 │ -0.985 │
└──────────┴───────┴────────┴────────┴─────────────┘
Consider the following arrays, converted to matricies:
let $$ = require('./dist/fluent-data.server.js');
let community = $$([
{ marker: 'Applewood Park', x: 0, y: 0 },
{ marker: 'Orangewood School', x: 10, y: 0},
{ marker: 'Kiwitown Market', x: 1, y: 10 },
{ marker: `The Millers`, x: -5, y: 0 },
{ marker: 'The Romeros', x: 0, y: -5 },
{ marker: 'The Lees', x: 5, y: 5 },
{ marker: 'The Keitas', x: 5, y: 0 },
{ marker: 'The Lebedevs', x: 15, y: 5 }
]).matrix('x, y', 'marker');
let transformer = new $$.matrix([
[ 1, 0.4 ],
[ 0, Math.pow(3,0.5) / 2 ]
]);
The following exmaple transforms the community data so that the new positions of the park, school, and market form an equilateral triangle. Then it analyzes the eigen properties of the transformer matrix.
let eigen = transformer.eigen();
community
.transform(transformer)
.log(null, 'Equilateralized Community:', 1e-8);
console.log('\nTransformer Eigenvalues:', eigen.values);
eigen.vectors.log(null, '\nTransformer Eigenvectors:', 1e-8);
Equilateralized Community:
┌───────────────────┬────┬─────────────┐
│ │ x │ y │
├───────────────────┼────┼─────────────┤
│ Applewood Park │ 0 │ 0 │
│ Orangewood School │ 10 │ 0 │
│ Kiwitown Market │ 5 │ 8.66025404 │
│ The Millers │ -5 │ 0 │
│ The Romeros │ -2 │ -4.33012702 │
│ The Lees │ 7 │ 4.33012702 │
│ The Keitas │ 5 │ 0 │
│ The Lebedevs │ 17 │ 4.33012702 │
└───────────────────┴────┴─────────────┘
Transformer Eigenvalues: [ 1, 0.8660254 ]
Transformer Eigenvectors:
┌────┬────┬─────────────┐
│ │ c0 │ c1 │
├────┼────┼─────────────┤
│ r0 │ 1 │ -0.94822626 │
│ r1 │ 0 │ 0.31759558 │
└────┴────┴─────────────┘
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent