Einstein Analytics is a powerful tool but as you might have noticed it doesn’t work like standard reporting. Because of the unique way data is stored in form of inverted indexing you can change data on the fly. So unlike other tools, including Salesforce standard report builder, you can in Einstein Analytics allow the user to modify the way they see data using different filters and selectors.
If you know your audience, which I discussed in a previous blog post as a key element in designing amazing dashboard, then you are probably also able to anticipate the questions they will be asking. So why not make sure your dashboard can accommodate those questions using filters and other selection options?
The platform supports different widgets that will allow you to select or filter data as you please this includes lists, toggle, date and range. There are two ways you can power these selections and filters; steps that are powered by your dataset or static steps where you define a static value you can use in other steps. The dynamic selection only works on filters, where static steps can work as filters, groupings, measures, limit and order. If you want to get started on static steps have a look here. Regardless both ways are quite powerful for the user to modify the dashboard and seek answers to their questions. Just bear in mind that though they are powerful, giving the user too many options will most likely end up doing the opposite and confuse them. So once you have applied these selection and filter options you want to take a step back and tell yourself “less is more”. Perhaps trial your dashboard with its many options to a few users that are inexperienced in Einstein Analytics, get their opinion and make the changes. Personally, I am very careful with using the date widget as my experience has shown me a lot of users get confused by its many options. Now I mostly use the list widget for date filters as you will also see in my sample dashboard below.
So going back to the point of knowing your audience; if you know who they are then you can anticipate their questions and thereby how to answer those. As you will see with my examples below, selection and filter options are a brilliant way to achieve this. I have created a sample dashboard to illustrate how selections and filters can be used in a dashboard, I have done this using data on the San Francisco Bike Share [find the data here].
How to apply selections and filters
In my example, I might anticipate that the viewers are interested in seeing total trips for a specific period. Instead of having multiple number widgets for different periods I allow the viewer to define their timeframe by having list filters for “Year”, “Month” and “Day”. This not only frees up space on the dashboard compared to having three or more widgets with different time periods, but it is also more flexible as the viewer might not want to see trips predefined to “this month” but rather how the trips looked during a big event like Dreamforce. I’ve also anticipated that they might actually be interested in knowing if the trips are part of an annual subscription or just casual, so I’ve allowed for that selection as well in form of a toggle widget.
Another example is that I have anticipated that the viewers are not only interested in seeing the most chosen start and end stations. In fact, I know they are also interested in knowing the least chosen start and end stations as this might help me decide if we need to move some bikes around. So instead of having four graphs, I have created two and I let the viewer choose if they want to see “most used” or “least used” by switching between the two in the toggle. This piece is powered by a static step.
So as you can see, allowing the viewer to select the story they want to see with different selectors, the dashboard ends up becoming quite powerful. Hence, your users can be empowered by the selectors and help them make better decisions.
When creating these selections and filters just remember to indicate which data the selection will apply to; is it for the whole dashboard as with the date selection, for the container as with the trip type or the individual graph as with the most and least used stations? This is key in ensuring that the user experience of the dashboard is high.