Create and Configure a New Dataset
Prerequisites
In order to create and configure a new dataset, you need both data and metadata. The data must be provided in a CSV data file that is well-formed and does not contain any illegal characters. In particular, do not use the forward slash character (/) because it can break the load process and result in errors during scoring. The metadata can either be provided in a JSON file or inferred automatically from the CSV data.
Overview
In Analytics Builder, creating a new dataset involves assigning a new dataset name, loading the set of data records that will be contained in the new dataset, and providing the metadata skeleton that will be used to structure the dataset. The metadata can either be uploaded as a pre-created JSON file or it can be inferred automatically from the CSV data itself.
 |
The following video demonstrates how metadata can be automatically inferred from the CSV file during dataset creation. It can be downloaded from the Reference Documents section of the PTC eSupport Portal:
|
To Create a Dataset
1. Click DATA in the left navigation panel to make sure the Datasets list page is open.
2. At the top of the Datasets page, click New. The New Dataset dialog box opens.
3. In the Dataset Name field, enter a name for the new dataset (preferably with no spaces).
4. For the Dataset Data file, click Choose File and navigate to a stored CSV file containing the data to be uploaded. The data file must include column headers.
5. To provide metadata, configure the following parameters:
◦ Upload metadata – Select this option if you want to upload a pre-created JSON metadata file. Then click Choose File and navigate to a stored JSON file containing the configuration information.
Clear this option if you want ThingWorx Analytics to infer the metadata automatically from the CSV data. The inferred metadata will be presented automatically for review and editing before the dataset is created.
◦ Review uploaded metadata – If you are uploading a JSON metadata file, select this option to display the metadata for review and editing before the dataset is created. This option is available only when Upload metadata is selected.
|
|
When metadata is inferred from the CSV data, it is automatically displayed for review and edit, so this option is not necessary then.
|
◦ Time series data – Select this option if the new dataset represents time series information. This option is available only if you will be reviewing the metadata.
6. Click Submit to begin the upload.
Status messages will be displayed in the Status Message field, the dialog box will close, and one of the following will happen:
◦ If you uploaded a JSON metadata file and did not opt to review the uploaded metadata, the new dataset is created and the Datasets list page is redisplayed with the new dataset in the list.
◦ If you either uploaded a JSON file and opted to review the uploaded metadata, or you opted to let ThingWorx Analytics infer the metadata automatically, a Review Metadata dialog box opens. For information about reviewing the metadata, see Review and Edit Metadata.
7. When you complete the review of the metadata, click Create Dataset. The new dataset is created and the Datasets list page is redisplayed with the new dataset in the list.