As COVID-19 continues to spread, healthcare groups and companies of all kinds are under pressure to provide care in the face of increasing demand. Healthy Data is your window into how data can help organizations address this crisis.
Data is everywhere and every company is focused on what to do with their available data stores: customer journey follow-ups, harvesting precious behavioral insights, performing market analyses, and uncovering sales predictions through Machine Learning, just to name a few. Analytics techniques can help companies beat their competitors, but what else can they do?
The whole world is focused on the coronavirus pandemic, which is responsible for shutting down borders, grounding planes, and imposing a world-wide lockdown. Companies and governments are turning to analytics to navigate these troubled waters.
There are many ways analytics and data will help the world overcome this crisis. In this article, we’ll focus on using Social Network Analysis to help companies understand their workforce and respond to coronavirus cases that hit close to home.
Defining a social network
Despite the popularity of David Fincher’s 2010 movie about the birth of Facebook, social networks are often misunderstood: they’re not “only” about your online social media. In fact, social networks have a long history, having been invented when mathematicians decided to impress sociologists with graphs:
These graphs are a representation of entities (each is assigned to a node) and their relationships (each relationship is represented by a line between two nodes). In the graph above, we can see that there are 12 entities with several connections between them.
There are a lot of theoretical studies about these graphs and they can be applied to solve many problems (classic examples are transportation problems), but here our focus is on a very specific type of graphs: those which connect people and are often associated with the study of social matters. So for our purposes, social networks are graphs whose nodes represent people and lines represent relationships between people.
These relationships can be obvious or not (two family members vs. two unrelated individuals who happen to like the same movie) and the information on each person can be abundant or scarce. Knowing this, what if a company possessed certain information on employees — home address, type of commute, team coworkers, desk location, etc. — were suddenly facing a pandemic? How could social networks analysis help?
Setting up your company’s social network
Social network analysis can provide useful insights into a company’s operations, help optimize internal processes, and even deal with human resources problems. But they can do more: they can help companies deal with a crisis like the outbreak of COVID-19.
Trying to tackle a pandemic caused by an infectious disease often involves some form of “social distancing,” but while some companies can put their workers on remote work, others simply can’t stop: power utilities, the foodservice industry, healthcare, etc. These jobs are critical to society, but the people doing them can also get infected by or infect their coworkers, which could prove disastrous because an uncontrolled situation like this could force these companies to shut down completely.
However, the management team can get help from analytics, setting up a social network of the company’s critical teams and creating a plan to contain the spread of the disease, keeping their business running as close to normal as possible.
Let’s look at how to do this:
Start by building a point-based system which focuses on a “proximity measure” of your workers. This proximity measure can be related to physical distance between coworkers, how they commute (crowded subways can get more points than a single driver on his own car), family relationships (siblings, parents, partners working for the same company), etc. The next table presents an example made with dummy data with information for each node.
Based on that points system, build a line-weighted graph and add as much information on each person (node) as possible: city they live in (for instance, if city A faces an outbreak, it will be crucial to know every employee who lives there), their age, etc.
And the result will be something like this:
In the graph, line thickness is related to the strength of relationships (based on the point-based system outlined above, which tends to group employees who share the same shift) and the colors of the nodes indicate the city where each employee lives in.
As you can see, there’s tons of information visually condensed into a small graph. We’ll use that information to make smarter decisions!
There’s a COVID-19 confirmed case. What can your company do?
Simple! Just cut the graph!
There are several levels of advanced techniques you can use to cut a graph (including using algorithms), but let’s keep it simple since a company’s critical teams aren’t often that big (or could be put into different shifts).
So, on the previous graph you can see that (unfortunately) Lynda contracted the disease. Given that, it’s easy to conclude that we can cut the graph by quarantining all of Lynda’s close coworkers. That is, all the people that have strong relationships (close “proximity” with Lynda should be sent home: Martha, Chris, and David. That would leave the company with two entire teams available and with low risk of contagion.
But this was an easy case.
Now, what if Lynda started to share her apartment with Anna, because rent was too expensive and the city she lived in too far away? Well, in that case, things could evolve in a very different manner.
Now, the problem is bigger.
And just by observing the updated graph, it’s easy to conclude that now 8 people need to be quarantined for precaution, leaving the critical team reduced to a third of its strength.
Using social network analysis to weather challenging times
Social Network Analysis is a powerful tool for companies.
As mentioned in this article, it can help keep the businesses running even in troubled times, specifically when a pandemic like coronavirus hits. However, that is not the only case.
A lot of companies are focusing on using this tool to help them fit the right people in the right job. That is the power of People Analytics: it helps, for example, companies find connectors, conflict solvers, or influencers — employees who are central in the company’s social network, who possess access to many colleagues or who can build bridges between very distant teams — among its workforce and give them a role that is in accordance to their particular set of skills.
So, Social Network Analysis is something essential to bear in mind for companies that look for modern ways to tackle their problems in a world where data is everywhere.
Ricardo Gonçalves has worked with data for more than eight years and is currently dedicated to applying Machine Learning techniques to predict power load and generation for the utility company he works for. He is also dedicated to empowering data teams to achieve excellency in data operations.