Need Help?

How to format your data

Last Updated: Mar 26, 2018
Data files must be in a specific format in order to be uploaded into Luminoso Analytics.  Before you can upload data to your project, please ensure that:
  1. Your spreadsheet's column headers follow the Luminoso Analytics guidelines (outlined below).  Data columns without correct headers will not be processed. 
  2. Your spreadsheet is saved in the comma-separated value (CSV) format.

Analytics v5 accepts the following kinds of data and metadata:  

Text data is the only required data column in any Luminoso project.  Label the column of the text data you want to analyze with the "text" header.   Analytics accepts only one text column per spreadsheet.  Individual rows may have blank text entries. 

You can add a title or identifier to each piece of text data.  Label the column containing the title data with the "title" header.  The title will be viewable but not analyzed with the text data.  

Label your categorical metadata with the column header “string_[FIELD]”.  The FIELD designates the metadata group (e.g. “string_Country of Origin”).  You may have multiple string columns, with as many values as needed, but will only be able to filter on fields with up to 10,000 values.  Note that fields with fewer than 30 values are represented as a list of checkboxes ...

... while fields with 30 - 10,000 values are represented as search boxes:

Label your numeric metadata with the column header “number_[FIELD]”.  Numeric metadata values can be either integers or floating point numbers.  The FIELD designates the numeric metadata by group (e.g. “number_NPS” or “number_Customer Age”).  Numeric metadata is filtered by specifying a range in the filter tool:

If you have relatively few discrete values (e.g. star ratings, NPS scores) and would prefer your numeric data to be represented as a list of checkboxes, you can label them as string metadata (see above).  

Analyze trends by adding one or more columns of date metadata to your project, which can be filtered by specifying a range in the filter tool:

We support all ISO 8601 compatible date and time strings, as well as Unix timestamps and US-style dates.

The following are some examples of ISO 8601 dates for April 10th, 2018:

  • 2018-04-10
  • ‚Äč2018-04-10 13:45
  • 2018-04-10T13:45:00Z

When using US-style dates, we support one or two digit months and days, as well as two or four digit years. Time can be specified as “HH:mm” or “HH:mm:ss” if seconds are desired. You’ll need to specify AM or PM (case insensitive) if using a 12-hour clock, as in “1:24 PM”.

The following are some examples of US-style dates for April 10th, 2018:

  • 04/10/2018
  • 4/10/2018 13:45
  • 4/10/18 1:45 PM

Note: All dates and times are assumed to be relative to UTC unless they are ISO 8601 strings, which can include a timezone designation.

A sample csv file

More Support
seconds ago
a minute ago
minutes ago
an hour ago
hours ago
a day ago
days ago
Invalid characters found