Indicator element: counts, sums and reference values

One Number at 2 A.M.. It's two in the morning during a winter storm, and the outage manager opens the dashboard with one question. How many customers are without power right now? She doesn't want a table, and she doesn't want to zoom around a map counting red dots. She wants one big, clear number, and ideally a hint of whether things are getting better or worse than an hour ago. That big number comes from an indicator element. If it's set up carelessly, it shows a confident number that answers the wrong question. In this lesson we'll make sure yours answers the right one.

What You Will Learn. Here's what you'll be able to do by the end. First, add an indicator element to your dashboard and point it at a layer, such as water main breaks or service requests. Second, pick the right statistic, which means knowing when to count features, when to add up a field, and when an average or a maximum tells the story better. Third, add a reference value, a second number the indicator compares against, so viewers can instantly see whether today is normal, worse, or better. We'll work through real examples along the way.

What an Indicator Is. An indicator is the simplest element in ArcGIS Dashboards, and often the most looked at. It's a card that shows one headline number. The number is calculated from a layer in your dashboard's web map, which is the saved map your dashboard reads its layers from, or from a layer you pick directly. The card has three lines of text. The middle line usually holds the big number, while the top and bottom lines hold a title and a short explanation, like "customers without power" or "updated every five minutes". Because the number is calculated from the layer, nobody types it in by hand. It updates each time the dashboard refreshes its data, and the refresh interval you set controls how often that happens.

How the Number Is Made. Let's slow down and follow the number from start to finish. It begins with a layer, say every traffic incident reported today. Next comes an optional filter, a condition that lets only some features through, like a sieve that keeps incidents where the status is active. Then a statistic squeezes the remaining features into a single value. That value can be compared with a reference, and the finished result is drawn on the card. Every indicator you build follows this same chain, so when a number looks wrong, walk back up the chain one step at a time.

Counts: How Many Things. The most common indicator is a count. In the indicator's data settings, set the value type to Statistic and choose Count. Count simply tallies how many features pass your filter, one per feature, just like counting people in a queue. For a service requests layer, add a filter where Status equals Open, and the card now shows how many requests are still waiting for a crew. That one setting turns thousands of rows into a number a supervisor can read from across the room. Change the filter to Status equals Closed, and you get every closed request in the layer, not just today's. For today's finished work only, add a second condition, such as closed date is today.

Sums and Other Statistics. Sometimes the question isn't "how many things", but "how much". That's when you use Sum. A field is simply a column in the layer's attribute table, and Sum adds up the values in one numeric field. For utility outages, summing the customers out field gives total customers without power, which is what our manager actually asked. Average gives the typical value, like the mean response time in minutes. Minimum and Maximum give the smallest and largest value of a field. For example, Maximum on a gallons lost field gives the size of the worst water main break. If your field stores seconds but you want minutes, value conversion lets you multiply or divide by a factor.

Worked Example: Count and Sum. Here's what the indicator calculates behind the scenes, shown as a small table of five outages. The filter keeps only the Open rows, which leaves three. Counting those rows gives three open outages. Summing the customers out column for those three gives one thousand two hundred, plus three hundred fifty, plus eighty, which is one thousand six hundred thirty customers. The maximum is one thousand two hundred, the biggest single outage. Notice that the two closed outages, even the huge one, are ignored because the filter removed them first.

Reference Values. A number on its own can't say whether it's good or bad. Is sixty open requests a busy day, or a quiet one? A reference value answers that. In the indicator's reference settings, Previous value compares against the number from the last refresh, which is great for live storm response. Fixed value compares against a number you type in, such as a service target of fifty. Statistic compares against a second calculation, for example the count of all requests, open or closed. You then place the reference in the card's text using the word reference inside curly braces.

Worked Example: Difference and Change. Let's work one through. Yesterday there were forty eight open service requests, and today there are sixty, so forty eight is our reference and sixty is our value. The forty eight comes from a Statistic reference: the same count of open requests, filtered to yesterday's date. You could also type it in as a Fixed value. Previous value would not work here, because it compares with the last refresh, not with yesterday. To get the difference, subtract yesterday's number from today's, which gives twelve more requests. To get the percent change, divide that difference by the reference, forty eight. Twelve over forty eight is one quarter, or twenty five percent. The indicator does this arithmetic for you. In the bottom text, type value in curly braces, then the word up, then percent change in curly braces followed by a percent sign, so the card reads sixty, up twenty five percent.

Common Mistake: Count When You Meant Sum. The most common mistake is choosing Count when the question needs Sum. In our example, Count said three, and Sum said one thousand six hundred thirty. Both are correct, but they answer different questions, and an executive who reads "three" as customers affected will badly underestimate the storm. The second trap is a missing filter, which quietly includes closed or cancelled records. Protect yourself with a simple habit. Before you share, check the indicator against a number you already know, such as a quick count in the layer's attribute table, and make the top text say exactly what the number measures.

Recap. Let's recap. An indicator takes a layer, applies a filter, and uses a statistic to produce one headline number that refreshes with the data. Count answers "how many things", while Sum, Average, Minimum and Maximum answer "how much" questions about a numeric field. And a reference value, whether it's the previous value, a fixed target, or another statistic, turns a bare number into a signal your viewers can act on, like sixty requests, up twenty five percent. Get those three pieces right and your indicator answers the question people actually ask.

Your Exercise. Now it's your turn. Open the dashboard you created earlier and use a layer with incidents, such as service requests, outages, or traffic incidents. You can take one from your own organization or from Living Atlas, Esri's free library of ready-made maps and layers. Add a first indicator that counts only the open or active features, and give it a fixed value reference that acts as a target. Then add a second indicator that sums a numeric field, like customers affected or estimated cost. Label both clearly in the top text and check each against the attribute table. In the next lesson, we'll arrange these elements neatly with stacking, tabs and a sidebar.