Javazone Beer Analytics with ELK

Beer analytics with ELK: Elasticsearch Logstash and Kibana

Christoffer Vigworks at

http://blog.comperiosearch.com

babadofar

Enterprise searchSearch based applicationsBig dataELK - logsMedia monitoringText analytics

AgendaIntroduction ElasticsearchVinmonopoletLogstash Kibana demoSearch algorithmsKibana beer analytics use cases

X

2010 Elasticsearch launched

2012 Elasticsearch the company 2015 Elastic{ON} - company name -> Elastic

Norwegian government owned alcohol monopoly.Sells beverages above 4,7 % alcohol Open data http://www.vinmonopolet.no/artikkel/om-vinmonopolet/datadeling

Vinmonopolet CSV file

CSV -> ElasticsearchLogstash

Logstash config

Logstash input

Logstash filter

Logstash output

Elasticsearch output

Kibana - Discover - DEMO

Field stats

Bitterness in beer

Bitterness - ordered

Select columns

AgendaIntroduction ElasticsearchVinmonopoletLogstash Kibana demoSearch algorithmsKibana beer analytics use cases

X

X

X

X

X

X

Use cases - Kibana beer analytics engine

Boss is buyingAnnual company party.Your mission is to buy beer for everyone, on behalf of the company.Should be expensive.Should taste “normal”

A beer for everyone?

A beer for everyone?

SORTED BY Price pr Liter

Lucene expressions search with filter

Boss is buying 2: Pie charts

Dashboard: Combining charts and search

APPLIED FILTERS

Gourmand customer visitA customer with taste for the good stuff is visitingYour mission: Take him out to dinner and make sure he is satisfied!

beer - countries

Beer - countries - table

Beer - countries - Significant terms

significant terms

Belgian beer significant terms

Belgium

barley

yeast

fresh bread

raisin

Significant terms - background filter

acidic

fresh bread

cherries

raisin

oak

Discovery tab - search

Sort by rating

AgendaIntroduction ElasticsearchVinmonopoletLogstash Kibana demoSearch algorithmsKibana beer analytics use cases

X

X

X

X

X

X

X

ResourcesElastic https://www.elastic.co/downloads ( no m in co)

code and vagrant box for this talkhttps://github.com/babadofar/javazone-box(elasticsearch 2.0 beta, Kibana source)

… Thank you!