Skip to main content

COVID-19 business continuity - gaining insights from analytics

This blog explores how utilities can gain insights and visualize data to assist in making decisions to support their organization and community. Learn how to implement detailed analytical capabilities on data then create simple and easy dashboards for both internal and external stakeholders.

Note: This blog is a transcription of the webinar ‘Business Continuity - Gaining Insights from Analytics’.

This article is the third in a series of three. You can find the other articles here:


  • Introduction: Location enabled analytics
  • Demonstration: Demographic enrichment & infographics
  • Demonstration: Chime model & forecast planning
  • Demonstration: Demand analysis & response
  • Demonstration: Operational awareness & proximity tracking

One of the biggest challenges that organizations are facing around the world is how to gain insights from all the datasets and information available that surround COVID-19. How do you bring all these disparate datasets together? Datasets that are predominately mastered outside your organization’s authoritative data. And how can you make sense of it all to make meaningful decisions?

ArcGIS helps organizations by providing:

  • Mapping
  • Visualizations
  • Analysis
  • Data science
  • Data collection
  • Situational awareness through operational dashboards

In this blog, we will look at four key focus areas for insights and analytics. We will show how ArcGIS can provide a better understanding of:

  1. Customers
  2. Work plans
  3. Network impact
  4. Safety and compliance


A proactive customer engagement plan comes from understanding your customers’ needs and potential circumstances. To achieve this, we need data on:

  • Population
  • Age profiles
  • Economic analysis
  • Working from home
  • Vulnerability to COVID-19

Demonstration: Demographic Enrichment & Infographics

ArcGIS Business Analyst has over 15,000 datasets in multiple categories such as population, income, and marital status. All of this data is collected by census and is available for use. Utilities can leverage this data to help decision making when putting together their COVID-19 response.

In our demonstration analysis, we’re trying to understand where the most infected areas are in our service area. We know that our most affected population will be seniors. The map indicates a higher density of seniors according to colour; the darker the area, the more seniors. Based on this, we can choose to modify the routing of our fieldwork so that they avoid the riskiest areas while performing their daily routine. 

The next variable we want to look at is the percentage of people with no home internet access. The darker the colour, the fewer the number of people with home internet access. This is a useful variable because we can infer that people without access to home internet are not able to work remotely; therefore, they may be experiencing a higher degree of financial hardship due to COVID-19 and may have trouble paying their bills.

In Business Analyst, we can use variables from a database, or we can bring in layers from the ArcGIS Living Atlas and ArcGIS COVID-19 Resource Hub. Both sources give access to thousands of daily-updated datasets that provide you with information on COVID-19 demographic and economics.

For example, we can pull in an unemployment dataset that characterizes the negative impact that COVID-19 can have on employment. This index represents the average job loss expected in the US. The higher the score, the higher the degree to which job losses might be higher than average. 

The real power of Business Analyst comes into play when we start combining several variables together. Here, we combine the three variables representing vulnerability to COVID-19 and apply a filter to only show those areas with a higher than average score. We can then use this map to reroute service routes around higher risk areas to reduce staff exposure. We could also reach out to customers who live in these areas to offer payment plans to help them pay for services should they be experiencing financial difficulty.

We can also use this information to power infographics that summarize key data points. 


Using ArcGIS Business Analyst we are able to easily visualize overall COVID-19 impacts, align these to demographics, understand where customers were potentially undergoing financial hardships to due to economic impacts in their region and identify high risk where it may be better to defer work to reduce the risk of exposing those high-risk areas.

Better understand

  • Population
  • Age profiles
  • Economic analysis
  • Working from home
  • Vulnerability

Make informed decisions around

  • COVID-19 impact assessment
  • Customer impact assessment
  • Financial hardship
  • Safety for customers and staff

Work Plan

An optimized ‘work plan’ comes from understanding the duration and balancing the forecasted program of work with resources, priorities and impacts to customers. Here are a few items we will look at in detail:

  • CHIME model
  • Resource availability
  • Preventative work
  • Capital work
  • New work to support COVID-19

Demonstration: CHIME Model & Forecast Planning 

Many organizations are using statistical models to learn how to better respond to COVID-19 and utilities are no exception. Of particular interest is the COVID-19 Hospital Impact Model for Epidemics (CHIME). 

Developed by Predictive Healthcare at Penn Medicine, the CHIME model leverages SIR modeling to assist hospitals with capacity planning around COVID-19. However, it can also give us an idea of what to expect in terms of a larger outbreak and the associated social distancing requirements and a stay at home guidance. Output for models like these generates that familiar curve that we’ve seen on the news. 

This model has been implemented as a key processing tool in ArcGIS Pro, which allows us to take GIS data as input and visualize that information geographically. Below, we see hospitalization rate by county.

We can improve our model by bringing in additional datasets - for example, from Esri’s Living Atlas and other data providers, such as Definitive Healthcare as seen in the below example.

We can also look at the degree to which social distancing is practiced in specific areas as seen in the below example. Essentially, this uses anonymized data from mobile devices to measure the change in average distance traveled pre and post COVID-19 to determine a social distancing score per county.

Here, we’ve set the model 200 days out to predict out based on the current input. It should be noted that there is a high degree of uncertainty surrounding COVID-19 so any long term, simplified model should be treated as tentative and with extreme caution. The information shown above is simply an example and does not reflect the most up-to-date information in this area.

In the below example, we see summaries and projections of COVID-19 hospitalizations, ICU admissions and ventilator admissions. Using this data, utility operators can better plan for the year, continue to provide essential services while staying safe and continuing to comply with guidelines.

Diving deeper, we can leverage dashboards to help utilities plan for operations in the context of COVID-19. In the below example, we’re looking at the planned proposed funded capital improvement projects across the service territory. At the bottom right, we’ve broken it down by the planned start date for these projects. Some are starting at the peak of the curve. This is probably the type of work that requires multiple individuals working together and is, therefore, not the best type of work to undertake when trying to comply with social distancing guidelines. Further, these projects are potentially disruptive to people working from home and kids studying in virtual classrooms. As such, this is the type of work utilities may want to push forward in the year when the curve has leveled off and social distancing guidelines are reduced.

So, what type of work can be pushed forward or postponed and what can be carried out? For many utilities, preventive maintenance and other routine work is probably the most appropriate to conduct in and around the peak of the curve. In contrast, labour-intensive capital improvement projects might need to be postponed until after the curve has flattened out. The below operations dashboard shows all the preventative maintenance for our utility in 2020.  The bottom right shows the due dates of assignments, which is fairly distributed throughout the year with a spike occurring at the end of the summer months. This is the type of work that can often be completed individually as it involves going from their home to the worksite without the need to interact with customers or other field crews. This is optimal work to conduct during the peak of the outbreak.

To get an idea of how to reallocate these works using all available information, analytical tools like ArcGIS Insights can be used. In this simple workbook, we’ve brought in the familiar curve from the model as well as the new location of field crews based on their home location who are not working from a central shop. On the left-hand side, we’ve also pulled in some additional data from the previous dashboard.

And now we can look at the map and see now people are routing from their home location and where they are in relation to their work assignments.

From here, we can do some simple analysis in the map view. For example, we can determine which workers are within a one-mile fixed distance from the various preventative maintenance assignments.


Using these tools and methods, we can build out fully functional dashboards that utilities can use to optimize work plans and complete all necessary work while keeping staff and customers as safe as possible. 

Organizations need new ways to optimize their planning processes. If this is not managed there will be large peaks in workflows that will be difficult to resource later in the year. Further, undertaking large projects that may result in service disruptions will likely negatively impact customers who are working from home. Lastly, the tools and methods we outlined above help organizations keep both their customers and staff as safe as possible.


COVID-19 has impacted demand in many industries. To operate the ‘new network’ requires an understanding of how demand has changed and where new opportunities may exist. This involves the following elements:

  • Demand changes
  • Working from home
  • Internet services

Demonstration: Demand Analysis & Response

Let’s examine how hourly residential demand for electricity has shifted. As seen in the below example, residential demand is expected to rise due to remote working. In contrast, commercial demand is expected to be lower while people shelter in place.

The challenge here is to understand which individual circuits are serving residential and which are serving commercial customers and how to handle the shifting and increased loads without inviting major issues into the network.

ArcGIS Insights lets you combine analytical graphs with geographic representations to make more informed decisions. The map below shows you areas with a high percentage of people working from home, which we can anticipate as having a higher demand for electricity.

We can overlay this map with a medium voltage line to see which parts of the network are impacted. We can also bring in layers showing impacted transformers to this map.

Down at the building level, we study an important pattern that will take place in the summer: shifting demand to the high penetration of solar in this area. 

In the below map, we see a medium-voltage line coming through this transformer, which goes down to the customer meter. The red dots represent residential solar. In a typical summer these solar panels would generate a high amount of power during work hours. Conversely, the buildings they are associated with would demand less energy during work hours. As such, these solar houses would push that surplus energy back into the grid.

However, this is not a typical summer. As more people work remotely, we can expect the demand for electricity in this area to be higher and a lower amount of surplus energy generated from residential solar. What we need to understand is which transformers and which readers are impacted by this shift in demand so that we can plan accordingly. 


With more people working and studying from home than they would otherwise, the demand for telecom services has increased. We can use ArcGIS Insights to examine the intersection between people working and studying remotely, demand for internet services and bundled internet services.

With this information, we can then decide which areas will likely need enhanced service delivery. We can even go further to identify marketing opportunities in areas with high demand but comparatively fewer bundled services subscriptions.

Safety & Compliance

All utilities are safety first both for staff and customers. The COVID-19 pandemic is like nothing we’ve had to monitor or track in the past. As such, we need more insights to manage our safety obligations. Let’s look at the following scenario of tracking the locations and work assignments of two employees: Employee 1 and Employee 2. 

Employee 2 has been assigned jobs 1 to 4 and they are starting their shift early. Employee 1 is at the top and only has two small inspection jobs and won’t start their shift until much later in the day. Employee 2 begins their day and they have 4 work assignments. This work requires a specialized vehicle, which they need to go to the depot to pick up. During the day, Employee 2 begins to feel unwell and returns the vehicle back to the depot and then heads home. 

Employee 1 is notified of additional work they need to complete. Employee 1 completes their shift, completes the 2 assigned jobs they already had and then continues on to complete the additional jobs. Finally, they return the vehicle to the depot and head home.

In this scenario, there was no person to person proximity. Although both employees were at the same location, it was not at the same time. However, as shown by the vehicle tracking, both employees were in contact with the same asset. Therefore, it is possible to assume that Employee 1 was exposed in addition to potentially four customers that were visited. 

This simple example illustrates how person to person and person to asset tracking can assist you in proximity tracking and auditing in order to meet business continuity obligations. 

Demonstration: Operation Awareness & Proximity Tracking

COVID-19 requires additional scrutiny and auditing of staff health conditions and activities. Let’s look at how information coming in from check-ins as part of the Coronavirus Business Continuity Solution as well as data from mobile apps like Tracker for ArcGIS can be used for this purpose. 

Using a dashboard like the below, an operations manager can quickly see if employees have checked in with symptoms. In the below example, we can see that one person has.

Knowing that a staff member has symptoms, we will want to know what that person has been doing over the past few days and where they may have been exposed. Using a dashboard like the below, we can do just that. Here we can see the location of field crews and the location of vehicles.

At this point, we want to focus on that individual who has been reporting symptoms. We can use their Survey123 check-ins to see if they have reported coming into close contact with anybody over the past few days and filter by only those locations in which they have.

We see that one employee did indeed come into close contact with another person. At this point, we’ll want to review what the employee was doing prior to being symptomatic. We can add a few more layers, including completed work orders and tracker tracks. This allows us to see the type of work they were performing and which customers they were in proximity to. We can then prepare the customer service department to reach out to any customers who may have been in contact with the symptomatic employee. 

We can further review the tracker tracks to determine if the symptomatic individual came into contact with other staff members. As seen in the below map, we can see that although this person was working in fairly close proximity to another staff member, they didn’t come into actual contact with that staff member.

We can even go one step further and use the historical AVL tracks, which reveals that although the two employees did not come into contact with one another, they did both use the same vehicle. We can, therefore, consider the second employee to have been potentially exposed.

Using these tools and the learnings derived from them, we can not only help identify who may have been exposed to COVID-19 but also determine and justify operating procedures that can be used going forward. For example, cleaning and disinfecting vehicles after use.

You can further explore Esri Canada solutions via the links below: