All Collections
Connect to our Datasets API
Create a new dataset in Geckoboard
Create a new dataset in Geckoboard

How to create a new dataset with defined schema fields.

Updated over a week ago

In this article, we'll outline the steps needed to create a new dataset with a defined schema in Geckoboard. This dataset can then be used with products like Zapier, Make and Parabola to post data to your dashboard.

Step 1: Add a new dataset widget

  1. Click Add widget, located in the top right of your dashboard.

  2. Search for Datasets using the Search sources field, then click on the Datasets data source button to open the connection panel.

    Search for datasets in our list of data sources
  3. At the bottom of the connection panel click on Create new dataset.

  4. Before we define the structure of our dataset first give it a name. Make sure to keep to lowercase letters, digits and no spaces.

    Naming a new dataset in Geckoboard

Step 2: Define your dataset schema

Now we'll add the schema fields that describe the columns that will make up our dataset. In our example the schema fields will correspond to the columns in an Airtable that records the sales of apples and pears and their shipping date.

Airtable spreadsheet showing sales of apples and pears

This table contains three data types: string, currency and date.

Airtable spreadsheet annotated with the different field data types

In total, there are 7 different types of fields you can add to your dataset: string, number, money, percentage, date, datetime, and duration. This table explains more about what data and format each field type can accept.

Field type




Use string fields for any sequence of characters. All string fields must not contain more than 256 characters.


Use number fields for integers and decimal values (e.g. 10.24). Use a string field if you want to add numeric fields to your unique key, or when using other characters like dashes - and brackets () (e.g. telephone numbers (555) 555-1234).


Use money fields for integers which specify the amount of money in a single currency’s smallest denomination (e.g. the USD’s smallest denomination is the cent, so a USD field would specify $10.00 as 1000).

You can specify the currency code when defining the field.


Use percentage fields for integers and decimal values. An integer in the 0 to 1 range will be displayed in the 0 to 100% range. A decimal value like 0.35 will be interpreted as 35% Values above 1 will correspond to percentages higher than 100% (e.g. 1.5 will be interpreted as 150%).


Use date fields when dates are formatted as YYYY-MM-DD (e.g. 2018-01-01). For hours, minutes and seconds, use the datetime field.


Use datetime fields when formatted according to ISO 8601. We recommend using the YYYY-MM-DDThh:mm:ssTZD variation, which will produce values like 2021-01-01T12:00:30Z (1st January, 2021, 12:00:30 pm, UTC).


Use duration fields for integers and decimal values. For example, setting your field to minutes and sending 1.5, will display as 1m 30s.

Can be set to milliseconds, seconds, minutes or hours.

To add your new dataset fields, follow these steps:

  1. Click Add new field. As you name each field its ID will be automatically generated.

    If you require empty/null fields, tick the Allow to be empty checkbox. If one or more fields make up a unique key, like our Order ID field, tick the Add to unique key checkbox. For more information, see our section on using empty/null and unique key fields.

    Also consider the type of visualization you'd like to build with your dataset. Some visualization types require specific field types. For more information, see our section on datasets visualization requirements.

    Adding the dataset schema fields in Geckoboard
  2. When you've added and configured all the fields for your dataset, click Create dataset.

  3. With your dataset created, you can now add data to it using products like Zapier, Make and Parabola, before creating visualizations from your dataset.

Adding empty/null and unique key fields

When adding certain field types to your dataset, you'll see additional options for allowing the field to be left empty or adding the field to a unique key.

Optional settings for new datasets fields in Geckoboard

In this example dataset we're showing customer records under four columns: Customer ID, Name, Town/City and Age. Here we can allow data in the Age fields to be empty/null and add data in the Customer ID fields to a unique key.

Airtable spreadsheet annotated showing list of customers
Airtable spreadsheet annotated with the unique order id and empty columns

Allow to be empty/null

There are cases where it makes sense for some columns or fields to not contain a value. In our example, Saskia, Lauren and Michael decide not to give their age. It makes more sense to keep those fields empty rather than using a zero '0', particularly when, say, calculating the average age of your customers. Those zeros will skew any results.

Your dataset must contain at least two fields for one to be allowed to be empty.

Adding to unique key

A unique key is a set of one or more columns or fields of a dataset that can uniquely identify a record in the dataset. You can use this key to create a unique identifier for easier retrieval later. Number fields cannot be added to a unique key, so we suggest a string field instead.

In our example, entries in Customer ID column can be unique because each customer must have a unique ID number.

Did this answer your question?