Early Access: Code-Free Data Exploration
Over the course of the next two months we'll be rolling out new features that turn every Mode report into lightweight B.I. Analysts will produce work in Mode the exact same way they do now, using the easiest, most powerful toolset available: SQL and Python.
Yesterday, this work produced static reports. Moving forward, every report is a powerful BI tool in its own right, granting coworkers the ability to drill down, segment, aggregate—and most importantly, share ideas—without having to write a single line of code.
An overview of can be found here: Code-Free Data Exploration.
There will be four phases of the rollout:
- Drag-and-drop charts
- Drill downs
- Advanced sharing
We'll be posting updates in this thread. Starting today we're rolling out Filters to a small group of customers, and will be onboarding additional folks before targeting general availability in the next two weeks.
We're beginning to roll out Filters, Phase 1 of Code-Free Data Exploration, starting today.
Here's a quick overview of how Filters work. Additional documentation can be found here.
Filters let you quickly hone in on a subset of a dashboard's data, without loading the database by re-running underlying queries.
After writing a query or two and building some charts, head over to the Report Builder and click Add Filter.
You'll find a list of column names of all the query results in your report. Select a column name to use as a Filter. You can then pick the Filter type and set a default.
Using Column Aliases
Filters become extremely powerful when you use column aliases in your queries.
Let's say you're building a high-level performance dashboard. You've got a chart of new users, powered by one query where dates come from a column called
activated_at.And another chart of revenue, powered by a query where dates come from a column called
In the Add Filter, you'd see both “Activated At” and “Date” - to abstract away the database structure and provide just one date range filter, simply alias one of the columns, like
activated_at AS date. Upon doing so, you can pick “Date” from the Add Filter menu and it will apply to both charts.
What about Parameters?
Using Parameters, a report viewer can modify the data that gets loaded into a report and require you to re-run the report's queries. Filters, on the other hand, impact the data once it's in the report and don't hit your database. In most cases, you'll want to use Filters instead of Parameters because they're faster to build, and make it easier for viewers to quickly iterate through ideas. When your data set is massive (like an event stream), a parameter can be used to load a reasonable amount of data into the report. Note: Reports can have both parameters and filters.
Filters are designed to work seamlessly with White-Label Embeds. Give exploratory power to customers right in your app or portal. Combined with custom themes, launched two weeks ago, delivering branded, interactive embedded analytics experience is easier than ever.
Donna McGahan For example in my report I have grouping types : gender, race and etc... I set Gender to default and it's nice for a client to filter between different grouping types and see different charts , but if they click "ALL" they get crazy looking graphs and outputs as it was not designed for customer to select ALL. I also do not want for some clients to access some of the filters, let's say data about state in New York. But it does not give me an option to remove some of the values in the filter.Reply
Hi Mans ! Did you run the queries in your report before looking at the filters?If you did run the queries and are still seeing a blank filter screen, could you please click the ? in the upper right corner of Mode and choose Contact Us to start a conversation with our support team? I'd like to get a link to your report where you're getting a blank filter screen. Thanks!Reply
Ana Robakidze , I've submitted a feature request to disallow All on a report filter. As for restricting what clients have access to which data, perhaps using a White-Label Embed might be a better option? https://help.modeanalytics.com/articles/setting-up-white-label-embeds/Reply