Interactivity

At first glance you may think that Visivo's push based architecture removes the ability to create interativity elements in your dashboard. However you can create powerful & dynamic interactivity in your dashboards by utilizing the selector.

Define a Selector

The selector enables you to toggle between trace cohorts with single select or multi-select. The benefit of this approach is that you still get deterministic builds that you can test while also being able to create drill down, granularity options, filtering ect.

Tip

You can leverage jinja2 loops and macros to easily generate multiple similar trace cohorts & charts.

For example you might want to create multiple different traces for different date grains and then add those traces to a single chart.

Example

# {% set date_grains = ['month', 'day', 'week', 'quarter', 'year']%} 
traces:
# {% for date_grain in date_grains %}
  - name: issues-created-per-{{date_grain}}
    model: ref(issues)
    cohort_on: "'{{date_grain}}'"
    props: 
      type: bar
      x: query(date_trunc('{{date_grain}}', created_at)::date::varchar)
      y: query(count(*) ) 
    order_by: 
      - query(date_trunc('{{date_grain}}', created_at)::date asc)
#{% endfor %}   
charts:
  - name: issues-created-total-over-time
    selector: 
      name: issue_metrics_date_grain
      type: single
    traces:
     #{% for date_grain in date_grains %}
     - ref(issues-created-per-{{date_grain}})
     #{% endfor %}   
    layout: 
      title: 
        text: Issues Created
      barmode: stack 
      legend: 
        orientation: h
        valign: bottom

Importance of trace.cohort_on

The cohort_on block of the trace enables you to dynamically split out series based into distinct groups. This is super helpful for dynamcially creating multiple lines, bars, ect with a single trace.

Example

Say you have a trace that shows revenue per week defined like this:

traces: 
  - name: revenue-per-week
    model: ref(orders)
    props:
      type: scatter
      x: query( date_trunc('week', date) ) 
      y: query( sum(amount) )

This would produce a single line with the week on the x axis and the sum of order amounts on the y axis. However, let's say you want to split this out by product type sold. To do that you can add a cohort_on attribute.

traces: 
  - name: revenue-per-week
    model: ref(orders)
    cohort_on: product_name
    props:
      type: scatter
      x: query( date_trunc('week', date) ) 
      y: query( sum(amount) )

These cohort values are also the primary way you can create interactivity across your charts by telling a selector, when it should show different trace cohorts.

Re-use the Selector

You are able to reference selectors from across your project on multiple charts By name using the ref() function. This enables you to connect charts to a single selector.

Note

If the selector is defined within a chart, all of the trace cohort values from that chart will be used to set the options for the selector.

However if you define the selector at the top of your project you can include as many traces as you want to pull in a broader selection of trace cohorts

Position the Selector in a Dashboard

You have two options for showing your selector in the dashboard.

1. You can define your selector within a single chart. If you do this, your selector will be present on that charts pop down menu. This is a great option if you want to provide interactivity without adding too much noise to the main dashboard canvas.
   

2. You can place the selector in a dashboard item to have it occupy a dedicated spot on the dashboard canvas.