JSTOR ngrams

JSTOR ngrams are indexed and stored on our secure S3 Buckets. The documents are indexed by their document ids (doi’s) on a MySQL database as well as in a collection of CSV files.


The Jstor ngram data is stored in a collection of 500 files (each approximately 3.2GB). These files are named n000 - n499. The indexes contain the filename and the byte indices of each doi. The schema followed by the database is also followed by the csv file. Here are a few lines from the jstor database.

doi filename line byteFirst byteLast nNgrams nWords
10.1086/231208 n000 0 0 2038709 5 26172
10.1086/233855 n000 1 2038710 3517884 5 19995
10.1086/314323 n000 2 3517885 4371013 5 12690
10.1086/316637 n000 3 4371014 4616514 5 7770

To connect to the MySQL database, please contact the Administrators for access. The database is accessible at `klab.c3se0dtaabmj.us-west-2.rds.amazonaws.com`_

If you plan to use the csv index files those are accessible from within Kotta by specifying the filenames as inputs to jobs.


The jstor index database is accessible from within Kotta as well.

Storage bucket

The data is currently stored on S3 Storage : s3://klabjstor.

The indices for the files in csv file formats are stored under the same bucket and the filenames are: s3://klabjstor/i{000 - 499}.csv

Since the data is organized by the data file, it is rather difficult to search for a particular doi without reading all 500 csv files. For search functions, it would be more convenient to query to external database and create a list of data files required and read out specific data items.

The datafiles are here: s3://klabjstor/n{000 - 499} are the files that contain the ngram data.


This data is only accessible from within the network and by users with the requisite access privileges.

Sample Jobs