ploner / rgeckoboard Goto Github PK

The package RGeckoboard makes it easy to feed Geckoboard Dashboards with data and images from R. The following functionality is integrated into the package:

• Access to the Dataset API (create, delete, update a dataset).
• Push R data to the Custom Widget API (e.g. for Barchart, Geckometer, Monitoring widget).
• Push R data directly to the Geckoboard PUSH API.
• Create an image optimized for Geckoboard usage.

An image created with this package must be uploaded to an URL accessible from Geckoboard. This could be a location on a cloud service (e.g. Dropbox, Google Drive) or a custom host. For more information see https://support.geckoboard.com/hc/en-us/articles/203676478#using_the_image_widget.

The package can be installed directly from github (latest version):

require(devtools)
devtools::install_bitbucket("ploner/RGeckoboard")
require(RGeckoboard)

or from the package repository CRAN (latest submitted version):

install.packages("RGeckoboard")
require(RGeckoboard)

The package depends on

• jsonlite
• httr

In order to use the various methods of this package, we need to save the account data of the Geckoboard Project into an R object of class geckoboardAccount. The next examples all make use of this account object.

## Fill in here the API key found on 
## www.geckoboard.com -> 'top-right' -> Account -> API Key. 
> account = geckoCreateAccount("ProjectName", key="553c55...")
> class(account)
[1] "geckoboardAccount"

This sample code defines an R data frame with a date column and 2 custom numeric columns. To create the data set on Geckoboard a custom name must be added plus the types of the columns. More information on the allowed types can be found on https://developer.geckoboard.com/api-reference/curl/.

Keep in mind that geckoDatasetCreate only creates an empty dataset which means it only uses the column names of the parameter df but not the rows. Data is updated in the next step, using the method geckoDatasetReplace.

> data = data.frame(
    date=c("2018-03-01", "2018-03-02", "2018-03-03", "2018-03-04"), 
	KPI1=c(77, 78.1, 79, NA), 
	KPI2=c(78, 90, 91.2, 92)
)
> datasetName = "kpi.by_day"
> types = c("date", "number", "number")
> geckoDatasetCreate(data, account$apiKey, datasetName, types)
> geckoDatasetReplace(data, account$apiKey, datasetName, types)

Numeric columns are marked by default as optional. This allows to submit NA values, too. If a field is optional or not can be changed using the parameter optional in the geckoDatasetCreate call.

To remove an existing dataset simply call:

> geckoDatasetRemove(account$apiKey, datasetName)

The functions provided within this package make it easy to specify the right resolution of a picture to be rendered on a Geckoboard dashboard. To generate an image optimized for retina display and with a size of 2x2 simply use geckoImagePrepare:

> geckoImagePrepare("folder/filename.jpg", size = "2x2", retina=TRUE, adjustTitleRow=FALSE)
> plot(100+cumsum(rnorm(50)), type="l", col="red", lwd=3, ylab="Stock Price")
> geckoImageSave()

The parameter adjustTitleRow should be used if the user adds a title row within Geckoboard, hence lowering the actual height of the image. The persisted JPG image must be moved to a cloud storage as described in https://support.geckoboard.com/hc/en-us/articles/203676478-Using-the-image-widget.

Using geckoCustomWidgetUpdate, data from a data.frame can be pushed easily to Geckoboard and visualized by a Bar Chart. First create a new CustomWidget in Geckoboard, then copy the widget key to paste it into R.

## Fill in here the widget API key found when creating a new CustomWidget on www.geckoboard.com.
> widgetKey = "..."
> df = data.frame(date=as.Date("2018-01-01", "2018-02-01"), val1=1:2, val2=rnorm(2))
> geckoCustomWidgetUpdate(
  widgetType="barchart",
  data=list(
    y=df[, c("val1", "val2")], 
    x=df$date
  ), 
  apiKey=account$apiKey, 
  widgetKey=widgetKey
)

Using geckoCustomWidgetUpdate, it is easy to create or update a Geckoboard Monitoring widget. First create a new CustomWidget of type Monitoring in Geckoboard, then copy the widget key to paste it into R.

## Fill in here the widget API key found when creating a new CustomWidget on www.geckoboard.com.
> widgetKey = "..."
> df = data.frame(date=as.Date("2018-01-01", "2018-02-01"), val1=1:2, val2=rnorm(2))
> geckoCustomWidgetUpdate(
  widgetType="monitoring",
  data=list(
    status="up",
    downTime="11 days ago"
  ), 
  apiKey=account$apiKey, 
  widgetKey=widgetKey
)

> data=list(...)
> geckoPushData(data, apiKey=account$apiKey, widgetKey=widgetKey)