1 Introduction

Chord plot includes 2 basic types:

Chord with Ribbon
Chord without Ribbon

The keys are:

data structure:
matrix mode: A data.frame comprising of a column of name, and a numeric matrix. Assign the name column to x, the matrix to y.
node/link mode: node data.frame [x, NA, series, weight]; link data.frame [x, x1, relation, value]. Combine them using rbind. If you don’t provide node data.frame, recharts will build it automatically. Assign the param list [x, x1, series/relation, weight/value] accordingly.

2 Function Call

echartr(data, x, <y>, <series>, <t>, <type>, <subtype>)

Arg	Requirement
data	source data in the form of data.frame
x	character independent variable. Other type will be coerced to factors. For node/link mode, you must provide two columns of `x`. For matrix mode, only the first column of `x` is accepted.
y	numeric dependent variable. For node/link mode, only the first column of `y` is accepted. For matrix mode, all the columns of `y` are accepted.
series	series variable which will be coerced to factors. Each level of `series` is treated as a subsetting factor to produce separate pies. Only the first one is accepted if multiple variables are provided.
t	timeline variable which will be coerced to factors. Only the first one is accepted if multiple variables are provided.
type	‘chord’.
subtype	chord: c(“ribbon”,“scale”,“scaletext”,“rotatelab”,“hidelab”,“clock”,“clockwise”,“asc”,“desc”,“ascsub”,“descsub”) ribbon: chort chart with ribbon. It requires dual-direction of links. scale: Show scale on the ribbon. scaletext: Show scale and text on the ribbon. rotatelab: rotate the tick label. hidelab: hide the label (default show). clock/clockwise: ascending order is clockwise (default anti-clockwise). asc: ascending nodes desc:descending nodes ascsub: asceding chords descsub: desceding chords

3 Showcase

3.1 Data Preparation

3.1.1 Matrix Mode

grpmtx <- matrix(c(11975, 5871, 8916, 2868, 1951, 10048, 2060, 6171, 8010, 16145,
                   8090, 8045, 1013, 990, 940, 6907), byrow=TRUE, nrow=4)
grpmtx <- as.data.frame(grpmtx)
names(grpmtx) <- paste0('Group', 1:4)
grpmtx$Name <- paste0('Group', 1:4)
knitr::kable(grpmtx, align=c('lllll'))

Group1	Group2	Group3	Group4	Name
11975	5871	8916	2868	Group1
1951	10048	2060	6171	Group2
8010	16145	8090	8045	Group3
1013	990	940	6907	Group4

The first 4 columns are exactly a matrix structure and the last column is a name vector. So this meets data structure requirements for matrix mode.

Matrix mode can be transformed to node/link mode as well. Matrix[i, j] represents 2 nodes (i & j) and 1 link (i -> j).

3.1.2 Node/link Mode

knitr::kable(deutsch[deutsch$year==2014,])

player	club	weight	role	year
Kruse	Monchengladbach	1	Fw	2014
Kramer	Monchengladbach	1	Mf	2014
Neuer	Bayern	1	Gk	2014
Boateng	Bayern	1	Df	2014
Lahm	Bayern	1	Df	2014
Kroos	Bayern	1	Mf	2014
Muller	Bayern	1	Mf	2014
Gotze	Bayern	1	Fw	2014
Badstuber	Bayern	1	Df	2014
Hummels	Dortmund	1	Df	2014
Weidenfeller	Dortmund	1	Gk	2014
Reus	Dortmund	1	Df	2014
Gundogan	Dortmund	1	Md	2014

There is not a numeric matrix structure inside the data.frame, so it is node/link mode. There are no NAs in column 1 or 2, so in this case, recharts will build a node data.frame based on the data.frame given, which looks like:

name	value	series
Badstuber	1
Boateng	1
Gotze	2
Gundogan	1
Howedes	1
Hummels	2
Kimmich	1
Kramer	1
Kroos	2
Kruse	1
Lahm	1
Leno	1
Meyer	1
Muller	2
Mustafi	1
Neuer	2
Ozil	1
Reus	1
Tah	1
Volland	1
Weidenfeller	1
Arsenal	2
Bayern	11
Dortmund	5
Leverkusen	3
Madrid	1
Monchengladbach	2
Schalke	2

3.2 Chord Chart

3.2.1 Chord with Ribbon

Set type ‘chord’ and subtype ‘ribbon’.

echartr(grpmtx, Name, c(Group1, Group2, Group3, Group4), 
        type='chord', subtype='ribbon + asc + descsub + hidelab + scaletext') %>% 
  setTitle('Test Data','From d3.js')

3.2.2 Chord without Ribbon

3.2.2.1 One-direction Relationship

Set type ‘chord’. If the data structure does not include dual-direction relationship, no matter whether you assign subtype ‘ribbon’, ribbon will not display properly.

echartr(deutsch[deutsch$year==2014,], c(club, player), weight, role, type='chord', 
        sub='asc + descsub + rotatelab') %>% 
  setTitle('Club Orientation of Deutsch Soccer Team (2014)')

3.2.2.2 Dual-direction Relationship

If you do want to display a chord with ribbon, you have to revise the data with dual-diretion relationship.

By now, deutsch14 only has one-direction relationship, which means, it only defines Gotze->Bayern, but not Bayern->Gotze.

knitr::kable(with(deutsch[deutsch$year==2014,], table(player, club)))

	Bayern	Dortmund	Monchengladbach
Badstuber	1	0	0
Boateng	1	0	0
Gotze	1	0	0
Gundogan	0	1	0
Hummels	0	1	0
Kramer	0	0	1
Kroos	1	0	0
Kruse	0	0	1
Lahm	1	0	0
Muller	1	0	0
Neuer	1	0	0
Reus	0	1	0
Weidenfeller	0	1	0

What we need to do is duplicate the source and target columns to build a matrix.

deutsch14 <- deutsch[deutsch$year==2014,]
deutsch14 <- rbind(deutsch14, deutsch14)
deutsch14[14:26, c('player', 'club')] <- recharts:::exchange(
    deutsch14[14:26, 'player'], deutsch14[14:26, 'club'])
knitr::kable(with(deutsch14, table(player, club)))

	Badstuber	Bayern	Boateng	Dortmund	Gotze	Gundogan	Hummels	Kramer	Kroos	Kruse	Lahm	Monchengladbach	Muller	Neuer	Reus	Weidenfeller
Badstuber	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Bayern	1	0	1	0	1	0	0	0	1	0	1	0	1	1	0	0
Boateng	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Dortmund	0	0	0	0	0	1	1	0	0	0	0	0	0	0	1	1
Gotze	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Gundogan	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0
Hummels	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0
Kramer	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0
Kroos	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Kruse	0	0	0	0	0	0	0	0	0	0	0	1	0	0	0	0
Lahm	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Monchengladbach	0	0	0	0	0	0	0	1	0	1	0	0	0	0	0	0
Muller	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Neuer	0	1	0	0	0	0	0	0	0	0	0	0	0	0	0	0
Reus	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0
Weidenfeller	0	0	0	1	0	0	0	0	0	0	0	0	0	0	0	0

In this way set subtype ‘ribbon’ and we get a chord with ribbon.

echartr(deutsch14, c(club, player), weight, role, type='chord', 
        sub='asc + descsub + rotatelab + ribbon') %>% 
  setTitle('Club Orientation of Deutsch Soccer Team (2014)')

Another example is based on preinstalled dataset mideast.

mideast <- as.data.frame(mideast, col.names=mideast[1,], stringsAsFactors=FALSE)
names(mideast) <- mideast[1,]
mideast <- mideast[2:16,]
me <- data.table::melt(mideast, id=NA)

## Warning in melt_dataframe(data, as.integer(id.ind - 1),
## as.integer(measure.ind - : '.Random.seed' is not an integer vector but of
## type 'NULL', so ignored

me <- me[!is.na(me$value),]
me$series <- strsplit(me$value, '/')
me$value <- sapply(me$series, function(x) as.numeric(x[2]))
me$series <- sapply(me$series, function(x) x[1])
names(me) <- c('source', 'target', 'value', 'series')
str(me)

## 'data.frame':    82 obs. of  4 variables:
##  $ source: chr  "叙利亚反对派" "阿萨德" "伊朗" "塞西" ...
##  $ target: Factor w/ 15 levels "美国","叙利亚反对派",..: 1 1 1 1 1 1 1 1 2 2 ...
##  $ value : num  1 1 1 1 1 1 1 1 9 9 ...
##  $ series: chr  "支持" "反对" "反对" "未表态" ...

echartr(me, c(source, target), value, series, type='chord', sub='ribbon')

3.2.3 Chord with Timeline

Let’s use year columns as timeline.

echartr(deutsch, c(club, player), weight, role, t=year, type='chord', 
        sub='asc + descsub + rotatelab') %>% 
    setTimeline(show=FALSE, autoPlay=TRUE) %>%
    setTitle('Club Orientation of Deutsch Soccer Team')

4 Futher Setup

Then you can configure the widgets, add markLines and/or markPoints, fortify the chart.

You can refer to related functions to play around on your own.

Basic Plots 15 - Chord Chart