Pivot is a web-based exploratory visualization UI for Druid built on top of Plywood.
The project is currently undergoing rapid development. Internal and external APIs may change with little notice.
Drag-and-drop UI
Contextual exploration
Comparisons
Make sure you have node (4.x.x) installed. On MacOS with homebrew you can do:
brew update
brew install node
Next simply run:
npm i -g imply-pivot
That's it. You are ready to Pivot.
Start off by running an example (static) dataset:
pivot --example wiki
Next connect Pivot to your broker by simply pointing it to your broker host
pivot --druid your.druid.broker.host:8082
Pivot will automatically introspect your Druid cluster and figure out your dimensions and measures.
Note: if Pivot starts up and gives you a query error it is most likely because it could not properly introspect your schema. You probably have some hyperUnique column that Pivot is trying to SUM over. You will have to provide Pivot with a config file as in the nest section.
In general Pivot will never know your schema as well as you. To get a better experience you should create a config and provide it to Pivot. The fastest way to create a config is to have Pivot do it for you.
pivot --druid your.druid.broker.host:8082 --print-config --with-comments > config.yaml
The --print-config
option will make Pivot run through its regular introspection and then, instead of spinning up a server, dump the YAML onto the stdout and exit.
pivot --config config.yaml
The next step is to examine and tweak the config using your favorite editor nano config.yaml
.
Here are the steps to clone Pivot and run it as a developer.
Firstly make sure you have gulp installed globally:
npm i -g gulp
Clone the project
git clone [email protected]:implydata/pivot.git
cd pivot
Inside the pivot folder run:
npm install
gulp
Finally you have to create a config.yaml
file. (or use the sample)
./bin/pivot --druid your.druid.broker.host:8082 --print-config --with-comments > config.yaml
The --with-comments
flag adds docs about what goes into the config.
Alternatively you can read the comments in the sample config file.
Then you are ready to
./bin/pivot --config config.yaml
Recent improvements:
- Better comparison behavior and legend interaction
- Support for query time lookups (ex.
$language.lookup('wikipedia-language-lookup')
) - Support for the extract function (ex.
resourceVersion: $resource.extract("(\d+\.\d+\.\d+)")
) - Custom dimensions (ex.
$user.substr(0,1)
) - Better legend for time series comparison
We will be working on:
- Additional visualizations (bar chart, geo, heatmap)
- Exclusion filters
- Better time selection
- String / RegExp filters
- Better time selection with date picker
- Various additions, improvements and fixes to make the app more complete
For updates about new and upcoming features follow @implydata on Twitter.
Please file bugs and feature requests by opening and issue on GitHub and direct all questions to our user groups.