This may be a strange question considering how much the term is used but I’m not sure it is truly understood. The tech buzz over the last several years has been data science, big data, and now artificial intelligence (AI). While these have importance, they have glossed over the eyes of most people in the industry, further separating what they need to do with how they do it. To help with this conversation, let’s reset the narrative and replace the term analytics with something business friendly like decision support. Decisions are what people make on a daily basis and the support they need is the topic of this article.

You would be hard pressed to find any data a utility manages where the value is something other than making decisions. The data managed in GIS is no exception. GIS data provides network, location, and attribution to some of the most important decision-making systems a utility uses – Outage Management, Advanced Distribution Management, and Power Engineering systems. These systems consume and leverage GIS data to create actionable information using advanced functionality – all so people can make decisions. This discussion, while focused on making decisions in general, is adding a GIS flavor – meaning looking at decision-making with a GIS infusion which may inspire you to pursue and capture the additional value.

Decision support has three high-level components. Each one of these components supports the ability to make decisions in a diverse utility environment. Most decisions are different and the context for which they need to be made are diverse. These component are:

Decision analysis outlines the way an individual in their everyday workflow makes decisions. At the lowest level a person needs to see something to give them the insight for what happens next. Not all decisions need a piece of software for an individual to do what they need to do, but sometimes they do. Of course it is important this individual have the right solution with proper tools to be efficient. There is no one-size-fits-all because that can be restrictive and technology is designed to enable workers, not impede them. By looking at different decision analysis capabilities, technology professionals can design systems and engineer data to be enablers. Decision analysis is broken down in the following areas:

Decision location specifies where decisions are made. This is an important factor because decisions are made in many places including the field, office, and in the board room. Applying a computing term to this, decisions are made at the edge or as we used to say at the right place, at the right time, by the right person. Defining this in the past was easy because technology was the limiting factor confined to heavy desktop applications managed by a specialized person and the results were not easily shared or presented. Today Esri’s platform has eliminated most of these limitations and makes where decisions need to be made a component of the solutions we as a community provide. Utility workers having actionable intelligence at the edge will increase the effectiveness of the decisions being made which is a positive direction and outcome.

Decision data is the data required for the type of analysis to be performed. Data is the most important piece to creating a good diverse decision-making framework which supports decision analysis and location – it is also the field where wars have raged. While this is not the topic for this discussion it is worth mentioning a rich analytical environment needs good authoritative data which may not have traditionally been managed in GIS – hence the war comment. There are two types of data sources for analytics, interactive and prepared. Interactive data is architected and part of a dataset that is managed for example, device status on a switch. The switch is either open or closed and I can see that in the data. The other type is prepared or engineered. This is data that has been specifically brought together, organized, and analyzed for a specific purpose and is presented in the form of an answer or awareness. Results are commonly found in dashboards or configured pop-ups but also in the form of another layer in a map. Without well thought out data architecture and engineered data, actionable intelligence is very hard to achieve.

Business decisions are the reason data is captured and maintained.  The directive is to manage authoritative data, leverage the powerful analytical capabilities, and develop good presentation frameworks for your users to use. I hope this focus on defining more about analytics has helped spark some ideas, and if it sparks questions reach out and we can discuss.

Let’s look at some examples and see if we can clarify some of the terms and ideas listed above.

Example 1: Decision Analysis – visual; Decision Location – field; Decision Data – Interactive

I am a field worker on the scene of a broken pole that needs to be replaced. I need to find the height and class of the pole and identify the attacher to inform them of the incident. I open my field device, find the pole on the map, and select the pole to see its attributes. From the data managed in my GIS I can see the height, class, and the attacher and proceed to get the materials needed and inform the attacher.

Example 2: Decision Analysis – Spatial; Decision Location – office; Decision Data – Prepared

I am an Engineer in the office focused on reliability planning. I need to see the worst performing circuits or areas, develop work orders to resolve, and submit a list of projects for budgeting. I open the reliability web application which already has my organizations outage events mapped, categorized based on outage root cause, and a very informative heat map drawing attention to clusters of bad areas.

Example 3: Decision Analysis – Network; Decision Location – office; Decision data – Interactive

I am an Engineering in the office investigating the recent increase in load on a particular circuit. I use GIS analysis tracing to calculate the connected load, which is at the meter or aggregated to the transformer, for the circuit. I then run scenario analyses by reconfiguring circuits by opening and closing switches then re-running the analysis tracing to see how the action impacted the load. This is a quick way to see the options before running power engineering models.