Newsletter

Three fundamental shifts with results

31 July 2015

The STEM results engine featured prominently in our recent article on STEM 7.5 development themes. We stated that graphs would be easier to draw, elements easier to select, and that result selections would be modifiable in situ.

In this short, follow-up article we expand on three specific enhancements in the pipeline which we expect to transform your everyday experience in both the desktop and online environments.

Drawing graphs, choosing what to graph

Almost every result in STEM is graphed automatically as a column or line chart according to its built-in characterisation as ‘aggregate’ (e.g., revenue and cost) or ‘instantaneous’ (e.g., connections, installed units). So the ‘how’ of graphing mostly looks after itself, whereas the ‘what’ can leave something to be desired when the lists of elements to select from become too large and cumbersome, and especially when using template replication on a large model.

The new results interface being developed for STEM 7.5 will simplify the selection of elements (and results and scenarios) in various ways:

  • all selection list-boxes will include a simple text filter to make it easier to home in on the element(s) you want
  • ‘two-stage selection’ will make picking non-contiguous items less fiddly
  • results will be classified out-of-the-box as Essential, Standard or Advanced, and you will be able to mark your own favourites and re-visit recently drawn results
  • template variants will be selected independently from the original elements (so you can select which elements first and then, e.g., which sites) with the option to select specific pairs as a ‘linear selection’ filtered first by element
  • when a model has multiple scenario dimensions, you will have the option to select scenarios in a ‘dimensional view’ as well as the conventional ‘linear selection’
  • new format options will allow you to hide any zero-value or unavailable series from the combinations selected for a graph (to be used with discretion).

Figure 1: Prototype tabbed selection dialog for drawing graphs in the forthcoming STEM 7.5

The new selection interface starts from elements rather than results (a theme which we will return to at the end of the article) and allows several new possibilities:

  • different types on the same graph; e.g., network and resource results together
  • ‘corner comparison’ against one specific base-element and scenario (or result), in addition to the current option only of an ‘edge comparison’ (limited to a base element or scenario, but not both).

Thus you will be able to draw specific example graphs such as:

  • service results as a proportion of revenue
  • resource costs as a proportion of the network total.

Note: it is helpful to think of the selections for a graph as a cube within an overall dataset of elements × results × scenarios, inferred by a selection of items in each dimension. The ‘corner comparison’ mentioned above relates to a corner of this cube. The ‘thin slice’ through this cube is explained in the following section.

Modifying a graph in situ

The current Results program allows you to draw and save graphs and tables in one or more views as part of a results workspace. However, the content for every new graph has to be selected from scratch. Possibly the single most valuable addition for STEM 7.5 is that you will be able to re-visit and modify the selections for a graph in situ, either to alter the graph, or as a template to make another one similar to it with some selections in common.

Understanding that interaction is a strong component of any presentation; we have also planned support for easy and intuitive ‘slicing’ between different elements, variants or scenarios, or indeed any ‘title element’ common to the selections on a graph or table:

  • a ‘thin slice’ presents results for just one element (or variant, scenario or whatever) but offers a user-determined subset of related elements via an accessible dropdown which makes it easier to rotate the graph between these different elements rather than reverting to the full model selection every time
  • a separate dialog interface allows you to choose (the inevitable extension of) a ‘thick slice’ combining more than one of the pre-determined elements on the same graph.

Figure 2: Rotating a chart between a user-determined subset of related elements

Immediate access to results in context

If there is one thing which none of the enhanced selection or re-selection logic addresses, it is the intrinsic connection between all of these results and the elements in the Editor to which they relate, and the rich set of calculations implied by their very existence.

So the next development milestone immediately following ‘a better results program’ is ‘no results program at all’! More specifically, our final ambition is to integrate the results access into the Editor itself so that this connection will be more apparent to the novice and more immediate for everyone.

Imagine selecting Results from the menu for a resource icon in the Editor and skipping the element selection altogether (one or more resources already selected), or even selecting an essential result like Installed Units from a cascading sub-menu directly. (This is the rationale for making the element selection first in the new selection interface shown above.) Optionally, the graph would show the same selection of scenarios as was most recently run from the Editor.

Figure 3: Graphing a result for a selection of resources in the Editor

Unlike the current implementation of graphs of inputs in the Editor, we envisage these new-style graphs as re-sizable, selectable objects which you could arrange in an Editor view just as you do in the current Results program, or even display alongside related icons and so on in a ‘normal view’. This opens up some rich possibilities for adding elements to graphs using the Connect tool or the familiar drag-and-drop interface.

Looking forward to the autumn

This work is still very much in progress, as it involved a significant re-working of the internal logic and data structures ‘behind the scenes’ before any of the user-facing benefits could be realised. The current intention is to demonstrate a beta system at the forthcoming STEM User Group Meeting in October 2015, and then to further refine this with feedback from the event towards an end-of-calendar year final release.

© Implied Logic Limited