To create a new dashboard
1. From the Home page, on the Dashboards tab, click the New Dashboard button.
2. In the New Dashboard menu that appears, you can select to use a template, but in this example, we select Blank Dashboard.
3. In the New Dashboard dialog, provide a Name for the dashboard. You can edit the Dashboard name and additional settings from the dashboard properties menu.
The additional settings available in the properties menu are:
- Tags - Enter tags to categorize dashboards for future use as templates.
When you enter a tag, it appears both on the tile for the dashboard, and in the list of Categories on the Templates page where you can click it to list all templates in that category. You can also use tags in the Categories box when searching for a dashboard.
- Dashboard Image - Upload a logo or other image, or select from the extensive library of Stratifyd images to represent the dashboard.
- Type - Change the Default value to Template to save the dashboard as a template that others can use.
- Description - Provide a description of the dashboard contents.
5. To add data, click the plus sign next to Data on the right side of the dashboard. In the Select a Data Stream dialog that appears, you can select an existing data stream, but here we create a new one.
6. In the New Data dialog that appears, you can upload a file, but here we select the Amazon data connector.
7. In the Data Connector Settings wizard that appears, supply the following information.
a. Specify the URL of the reviews page for the item and click Next. (To specify any additional Amazon review URLs, click Add new search.)
b. Add a comment limit
8. On the Schedule page of the wizard, you can opt to set up a recurring schedule to run the crawler periodically, but here we use the default one-time date range of one month. Click Submit.
9. On the Connect to the Data Stream dialog that appears, change the name of the data stream and click Submit.
To add a data model
Once the structured data analysis is complete, i.e. the Data Ingestion panel indicates that the Structured Data Analysis model is up to date, you can add data models to process your data.
1. In the Data panel, click the vertical ellipsis and select Edit.
2. In the Edit Data Streams dialog that appears, click Deploy Model.
3. In the Deploy Model dialog that appears, you can choose from supervised models, basic analyses, or any pre-trained models that may have already been created, but here we select the Unsupervised NLU Model.
4. In the Deploy a New Model wizard that appears, for the Amazon connector, assumed fields are already mapped to the Text and Dates fields in the Assigned Fields column.
You can click the red x to remove them, and click the blue + next to any fields in the Unassigned Fields column to map them, but here we accept the defaults and click Next.
5. On the Submit page of the wizard, you can optionally change the name of the model and add tags and a description, but here we click Submit.
6. In the Edit Data Streams dialog that appears, click Submit.
The model appears in the Data Ingestion panel in the processing section. When it finishes processing, it moves to the up-to-date section.
To add a widget
Start by creating a new blank tab to be your workspace area. To do so, click the plus button next to the data tab.
1. In the dashboard widgets workspace area, at the top right, click the Add Widget button. You can also right-click anywhere on the dashboard and paste or insert a widget at the mouse location.
2. In the Widget Editor that appears, in the data panel to the left, select a dimension on which to base the widget visualization.
You can select calculated fields or original fields from the structured data, but here we select the sentiment.overall dimension from the Unsupervised NLU Model.
3. The default visualization type for a sentiment analysis dimension is the gauge.
Other dimensions have different default visualization types. The Visualization tab of the panel to the right updates with valid selections for each type. A sentiment analysis dimension can be changed to any of the following types: bar, treemap, pie, funnel, bubble, waffle, scatterplot, or list.
You can change this to another type of visualization, but here we leave the default gauge. Click the title text for the widget and provide a title, for example, Overall Sentiment.
4. In the Data Dimensions tray along the top, you can further customize your widget:
- Drag in more data dimensions
- Drop down a selected dimension and set aggregation, format, bins, and type
- Add custom filters and notes
5. In addition to the Visualization tab, other tabs update with valid options for the selected visualization type.
- The Components tab lets you add or remove components of the widget and offers options that apply to each
- The Colors tab lets you set color styles and ranges
- The Annotations tab is available for bar, line and scatterplot visualizations
6. When you are satisfied with your settings, at the top of the screen, click Save.
7. The widget is added to your dashboard where you can resize it using the grab handle on the bottom right corner or move the widget by clicking and dragging any long edge of the widget.
8. Click the hamburger menu on the top right corner of the widget to edit, copy, or delete the widget, or add a memo or filters.