Datasets is where users can upload data to NeoPulse® Manager. Later, the uploaded data can be attached to projects for use in the training and querying of models (see Models). For now, Datasets has one component, the "Overview" tab, in which all of the functionality is contained. The main actions available there will be explained in the sections to follow.

Example: a Datasets Overview Tab

Creating Datasets

Datasets are imported to NeoPulse® Manager in a zip file format. Within the zip file, there must be a single CSV titled "index.csv". Media files can also be included in the zip file in whatever organization is desired. To demonstrate the proper file format for dataset upload, consider the following example of a valid zip structure:

Example: zip file containing image data in a folder called "dogs"


NOTE: The columns of index.csv should refer to the images in the dogs folder using a relative path.

Once the dataset zip file has been properly formatted, it can be uploaded to NeoPulse® Manager using the "Upload Dataset" button in the "Overview" tab. Upon pressing this button, a pop-up will give users the opportunity to attach a name, description, and tags to the new dataset.

Example: the Upload Dataset form

Viewing and Editing Datasets

After a dataset has been uploaded, it will appear in the Datasets Tab alongside any other datasets previously uploaded to NeoPulse® Manager. To view an uploaded dataset, simply click on it to see the contents of its "index.csv". Column headers will appear in bold with corresponding data displayed underneath. If multiple pages are needed to display all of the CSV's data, then these pages can be selected or scrolled through beneath the CSV's view. Exit the view by pressing the "Back" button at the top left of the screen.

Example: a Dataset's view

While viewing a dataset, users may also edit the dataset. Next to the dataset's title are two buttons: a pencil and a trash can. Selecting the pencil icon allows for editing of the dataset's name and description; the trash can icon prompts the user for confirmation of deletion of the dataset from NeoPulse® Manager. A dataset's tags can also be edited by clicking the "edit" button to the right of the currently displayed tags. Finally, while importing a dataset, NeoPulse® Manager automatically formats the column's data_type into one of the seven currently-supported types (see how these are used in an NML file's source construct). However, users may override the default data_type option by clicking the icon to the left of the column header, which creates a drop-down menu where the desired option can be selected.


In NeoPulse® Manager, tags are a means by which to sort datasets. Tags are best used as one-word descriptions of properties related to the data. While creating or editing a dataset, all currently-existing tags will can be added to the dataset from a drop-down menu. To create a new tag, simply type it into the Tags field and press enter when finished typing. Then while within the "Overview" tab, all currently-existing tags will be displayed above the datasets uploaded to NeoPulse® Manager. Toggling a tag will filter out any datasets who don't have the tag applied to them, allowing for easy sorting of datasets by keywords. Multiple tags can be toggled at once for further sorting.

Example: activating multiple Tags to filter out Datasets in the Overview Tab