Bime 1.55: Your Dashboard As A Slideshow [RELEASE]

Bime release 1.55 is out. Only 2 weeks after Bime 1.50, it's a classic example that good features are not the most time consuming to build. We love incremental improvements, we love SaaS.

And so, your dashboards become slideshows...

One of our many happy customers told us he wanted more in the dashboard department. His first point was that the zoom on visualizations was just not as good as it could be: "visualizations are not taking full advantage of their size". Indeed, the zoom was...well much like magnifier glass: reliable, dependable, but no longer pushing the boundaries of discovery. Now, in 1.55, the zoom produces a size smart visualization. Simply, if you have a visualization with a lot of data points and you zoom in, you'll find much higher chart definition with detailled axis, more labels, easier data tips and a whole lot more insight.

Our friendly correspondent was also using Bime's dashboards as one-stop slideshow shops. During board meetings he used an interactive dashboard view, backed by every last scrap of data he owned. This allowed him not only to present his main data but to answer questions on the fly and respond quickly to audience reactions. What a great idea! Our dev team had pushed the slideshow analogy to the max: when zoomed, you can now navigate easily through your charts and quickly dip into your database to find that key set of backup data. We never want to hear of another executive being caught short without data!

Time intelligence calculated members and measures

1.55 includes a lot of new time intelligence formulas. Here are some common use cases for your delectation:

  • Sales for 2007: calculated measures => IF(YEAR(Sales date)="2007") THEN SUM(turn over) ENDIF
  • Sales for last month: calculated measures => IF( MONTH(Sales date)=CURRENT_MONTH()-1 AND YEAR(sales date) = CURRENT_YEAR()) THEN SUM(turn over) ENDIF

Warning: Advanced feature. Disaggregated Data

Bime 1.55 introduces SUMX Agregation. "Why is it useful?" we hear you cry. Well, it allows you to apply a formula on each row and get the global result back. As always, an example's best:

Let's say you have 2 lines of data...

Unit price: 5, Quantity sold: 10

Unit price: 10, Quantity sold: 10

...and that you have a calculated measure that does:  SUM(Unit price) * SUM(Quantity sold)

With a typical SUM aggregator on the calculated measure the result would be: 300 for (5+10) * (10+10). Obviously, that is not what you wanted! If you instead apply a SUMX on the calculated measure, you'll get: 150 for (5*10) + (10*10). Much better.

Bime now evaluates the formulas on each row instead of on the aggregated data. The drawback is an increase in calculation density, so use only when necessary. In our example there is a strong relationship between the measures so it is definitely the way to go.

Fixes and minor improvements

Improvements: Great news! In Bime Desktop there are no more Popups for Déjà Vu synchronisation. Value filters in dashboards are applied locally rather then globaly; you can now change the color of your dashboard titles (in admin > account); you can now update Excel Datasource; refinement in complex TOP filters; better display of basket attributes; COUNT_VALUE & DCOUNT_VALUE functions for counting (distinct) attribute values; use of a constant for bullet graph budget; zoom level for heatmap is persisted; you can dynamicaly change the order of filters.

Fixes: display name replaces the technical name in every visualization; some old queries get an extra % in the value displayed; windows added to the dashboard inherit the query title; consistent Pie chart labels; Autobuild when the datasource has several column of the same geo level (eg. several cities) could produce bad schemas.