lineupjs v4.10.0

LineUp.js: Visual Analysis of Multi-Attribute Rankings

(dev)

LineUp is an interactive technique designed to create, visualize and explore rankings of items based on a set of heterogeneous attributes.

Key Features

• scalable (~100k rows)
• heterogenous attribute types (string, numerical, categorical, boolean, date)
• composite column types (weighted sum, min, max, mean, median, impose, nested, ...)
• array (multi value) and map column types (strings, stringMap, numbers, numberMap, ...)
• filtering capabilities
• hierarchical sorting (sort by more than one sorting criteria)
• hierarchical grouping (split rows in multiple separate groups)
• group aggregations (show a whole group as a single group row)
• numerous visualizations for summaries, cells, and group aggregations
• side panel for easy filtering and column management
• React, Polymer, RShiny, Juypter, and Power BI wrapper

Usage

Installation

npm install --save lineupjs@next

<link href="https://sgratzl.github.io/lineupjs_docs/master/LineUpJS.css" rel="stylesheet">
<script src="https://sgratzl.github.io/lineupjs_docs/master/LineUpJS.min.js"></script>

Minimal Usage Example

// generate some data
const arr = [];
const cats = ['c1', 'c2', 'c3'];
for (let i = 0; i < 100; ++i) {
  arr.push({
    a: Math.random() * 10,
    d: 'Row ' + i,
    cat: cats[Math.floor(Math.random() * 3)],
    cat2: cats[Math.floor(Math.random() * 3)]
  })
}

const lineup = LineUpJS.asLineUp(document.body, arr);

CodePen

Advanced Usage Example

// arr from before
const builder = LineUpJS.builder(arr);

// manually define columns
builder
  .column(LineUpJS.buildStringColumn('d').label('Label').width(100))
  .column(LineUpJS.buildCategoricalColumn('cat', cats).color('green'))
  .column(LineUpJS.buildCategoricalColumn('cat2', cats).color('blue'))
  .column(LineUpJS.buildNumberColumn('a', [0, 10]).color('blue'));

// and two rankings
const ranking = LineUpJS.buildRanking()
  .supportTypes()
  .allColumns() // add all columns
  .impose('a+cat', 'a', 'cat2'); // create composite column
  .groupBy('cat')
  .sortBy('a', 'desc')
  

builder
  .defaultRanking()
  .ranking(ranking);

const lineup = builder.build(document.body);

CodePen

Supported Browsers

• Chrome 64+ (best performance)
• Firefox 57+

• Edge 16+

React Support (LineUp.jsx)

A React wrapper is located at lineupjsx.

Installation

npm install --save lineupjsx

<link href="https://sgratzl.github.io/lineupjs_docs/masterx/LineUpJSx.css" rel="stylesheet">
<script src="https://sgratzl.github.io/lineupjs_docs/masterx/LineUpJSx.min.js"></script>

Minimal Usage Example

// generate some data
const arr = [];
const cats = ['c1', 'c2', 'c3'];
for (let i = 0; i < 100; ++i) {
  arr.push({
    a: Math.random() * 10,
    d: 'Row ' + i,
    cat: cats[Math.floor(Math.random() * 3)],
    cat2: cats[Math.floor(Math.random() * 3)]
  })
}

<LineUp data={arr}/>

CodePen

Result is same as the builder minimal example

Advanced Usage Example

// arr from before
<LineUp data={arr} defaultRanking>
  <LineUpStringColumnDesc column="d" label="Label" width={100} />
  <LineUpCategoricalColumnDesc column="cat" categories={cats} color="green" />
  <LineUpCategoricalColumnDesc column="cat2" categories={cats} color="blue" />
  <LineUpNumberColumnDesc column="a" domain={[0, 10]} color="blue" />

  <LineUpRanking groupBy="cat" sortBy="a:desc">
    <LineUpSupportColumn type="*" />
    <LineUpColumn column="*" />
    <LineUpImposeColumn label="a+cat" column="a" categeoricalColumn="cat2" />
  </LineUpRanking>
</LineUp>;

CodePen

Result is same as the builder advanced example

Polymer Support (LineUp-Element)

A Polymer 2.0 web component wrapper is located at lineup-element.

Installation

bower install https://github.com/sgratzl/lineup-element

<link rel="import" href="bower_components/lineup-element/lineup-element.html">

Minimal Usage Example

// generate some data
const arr = [];
const cats = ['c1', 'c2', 'c3'];
for (let i = 0; i < 100; ++i) {
  arr.push({
    a: Math.random() * 10,
    d: 'Row ' + i,
    cat: cats[Math.floor(Math.random() * 3)],
    cat2: cats[Math.floor(Math.random() * 3)]
  })
}
conat data = { arr, cats };

<lineup-element data="[[data.arr]]"></lineup-element>

TODO CodePen

Result is same as the builder minimal example

Advanced Usage Example

// arr from before
<lineup-element data="[[data.arr]]" side-panel side-panel-collapsed default-ranking="true">
  <lineup-string-desc column="d" label="Label" width="100" ></lineup-string-desc>
  <lineup-categorical-desc column="cat" categories="[[cats]]" color="green" ></lineup-categorical-desc>
  <lineup-categorical-desc column="cat2" categories="[[cats]]" color="blue" ></lineup-categorical-desc>
  <lineup-number-desc column="a" domain="[0, 10]" color="blue" ></lineup-number-desc>
  <lineup-ranking group-by="cat" sort-by="a:desc">
    <lineup-support-column type="*" ></lineup-support-column>
    <lineup-column column="*" ></lineup-column>
  </lineup-ranking>
</lineup-element>

TODO CodePen

Result is same as the builder advanced example

R, RShiny, and R Markdown Support

A HTMLWidget wrapper for R is located at lineup_htmlwidget. It can be used within standalone R Shiny apps or R Markdown files. Integrated plotting does not work due to an outdated integrated Webkit version in RStudio. Crosstalk is supported for synching selections and filtering among widgets.

Installation

devtools::install_github("rstudio/crosstalk")
devtools::install_github("sgratzl/lineup_htmlwidget")
library(lineup)

Examples

lineup(iris)

Jupyter Widget (to be released)

A Jupyter Widget wrapper for Python is located at lineup_widget.

Installation

pip install -e git+https://github.com/sgratzl/lineup_widget.git#egg=lineup_widget
jupyter nbextension enable --py [--sys-prefix|--user|--system] lineup_widget

Or, if you use jupyterlab:

pip install -e git+https://github.com/sgratzl/lineup_widget.git#egg=lineup_widget
jupyter labextension install @jupyter-widgets/jupyterlab-manager

Examples

import lineup_widget
import pandas as pd
import numpy as np

df = pd.DataFrame(np.random.randint(0,100,size=(100, 4)), columns=list('ABCD'))

w = lineup_widget.LineUpWidget(df)
w.on_selection_changed(lambda selection: print(selection))
w

from __future__ import print_function
from ipywidgets import interact, interactive, interact_manual

def selection_changed(selection):
    return df.iloc[selection]

interact(selection_changed, selection=lineup_widget.LineUpWidget(df));

PowerBI Custom Visual (under development)

A PowerBI Visual wrapper is located at lineup_powerbi.

Installation

TODO

Examples

TODO

API Documentation

See API documentation and Develop API documentation

Demos

See Demos, Develop Demos, and R Demos

Related Publications

LineUp: Visual Analysis of Multi-Attribute Rankings Paper Paper Website

Samuel Gratzl, Alexander Lex, Nils Gehlenborg, Hanspeter Pfister, and Marc Streit IEEE Transactions on Visualization and Computer Graphics (InfoVis '13), 19(12), pp. 2277–2286, doi:10.1109/TVCG.2013.173, 2013.

:trophy: IEEE VIS InfoVis 2013 Best Paper Award

Taggle: Scalable Visualization of Tabular Data through Aggregation Paper Preprint Paper Website

Katarina Furmanova, Samuel Gratzl, Holger Stitz, Thomas Zichner, Miroslava Jaresova, Martin Ennemoser, Alexander Lex, and Marc Streit arXiv preprint, 2017.

Dependencies

LineUp.js depends on

• LineUpEngine table rendering engine
• D3 utilities: scales, format, dragging
• Popper.js dialogs

Development Dependencies

Webpack is used as build tool. LineUp itself is written in TypeScript and SASS.

Development Environment

Installation

git clone https://github.com/sgratzl/lineupjs.git -b develop
cd lineupjs
npm install

Build distribution packages

npm run build

Run Linting

npm run lint

Serve integrated webserver

npm run start

Authors

• Samuel Gratzl (@sgratzl)
• Holger Stitz (@thinkh)
• The Caleydo Team (@caleydo)
• Datavisyn GmbH (@datavisyn)

This repository was created as part of the The Caleydo Project.