Awesome Open Source
Awesome Open Source

elastic datasets

This is a collection of smallish datasets to use for playing with Elasticsearch.

You can only fit so much data in an R package. The R client for Elasticsearch we maintain elastic comes with some data, but of course it's nice to have more, so here it is.

See also nodbi for working with Elasticsearch from R.


  • plos_everything.json
  • plos_introductions.json
  • plos_data.json
  • - too big for gitub, at dropbox
  • gbif_data.json
  • gbif_geo.json
  • gbif_geopoint.json
  • gbif_geoshape.json
  • gbif_geosmall.json
  • shakespeare_data.json
  • omdb.json

Loading into ES

These datasets are formatted to be ready for bulk loading into Elasticsearch via the bulk API

geonames is about 70 .json files in Elasticsearch bulk format. It was prepared from the Geonames database at The original data from Geonames was licensed under a Creative Commons Attribution 3.0 License, see

To load the geonames data into Elasticsearch, do as you wish, but e.g., in R you could do:

First, create the index and set the geo_shape mapping

body <- '{
 "mappings": {
   "record": {
     "properties": {
         "location" : {"type" : "geo_shape"}
index_create(index='geonames', body=body)

should return

#> $acknowledged
#> [1] TRUE

Note: the index type is record, and the index name is geonames. The index and index type were set in the json files.

Then use a for loop to load in each file. AKAIK there is a limit on the file size you can load in (let me know if there's a way to get around it), so that's why theres a bunch of json files instead of one big file.

files <- list.files("path/to/unzipped/files")
for(i in seq_along(files)){
      sprintf("path/geonames%s.json", files[i])

The docs_bulk() function uses the /_bulk endpoint to POST data to an index called geonames in your ES server. The output of the bulk load call prints info, that's why we use invisible() so you don't get thousands of lines printed.

Check that it worked:

#> [1] 6646030

You should have ~ 6.6 million records

Get A Weekly Email With Trending Projects For These Topics
No Spam. Unsubscribe easily at any time.
elasticsearch (556
dataset (422
elastic (54