I have been trying to do an aggregation using Kibana-4 and in the aggregations dropdown there is a new type of aggregation saying "siginificant trems" . Would someone explain what it is and what is the best way to visualise it?.
Significant aggregations is a type of aggregation which will return unusual or interesting occurrence of terms in a set of data. This is an experimental functionality as of now in elasticsearch,and this has been included in Kibana 4.0 too. To make things clear,let us do an area graph visualization on an index "comics". Here I have set the Y-axis to the count of occurrences,then an "split area",where I applied a range aggregation(ranges are from 0-10 and 10-30) and the X axis to have a sub-aggregation of "significant terms",with a size value equal to 4. And the resulting visualization is given below: You can see in the graph,that there are two area graphs and in the X axis,the hashtag values of our aggregations. Here for eg:,the term "australia" is listed as one of the unusual terms.This may be because of the rise in the tweets with hashtag Australia,as they have won the cricket world cup as recent as 2 days before. So,this aggregation helps in detecting anomalies or unusual patterns in the data set as illustrated above.