Responsibility with data

It’s easy to be a superhero when you have skills with data.

Being able to analyse data gives you super-vision, letting you see things you couldn’t see before. It can give insight into what might happen in the future – prescient knowledge. It gives super-speed because you can react and adapt to situations much more quickly.

But to quote Spider-Man: with great power comes great responsibility.

It’s all well and good to do amazing things with data. To see that a particular value is going up while another value isn’t. To improve data quality. To learn about how data relates to businesses. To develop a culture around data. To use data effectively.

It’s all well and good because if we focus too much on the numbers we can disconnect the data from what it means. In some ways, this can provide an objectivity that can become a strength. Letting the numbers speak for themselves without emotional ties. No strings.

Except that data is not just numbers. Data refers to things. To people. To loan applications. To medical emergencies. To salaries. Things that should never be considered just numbers. Reduced through objectivity to mere objects. These things are the subjects of life, and we need to always retain some amount of subjectivity.

It’s easy for an accountant to look at the cost of the people in an organisation, and decide that that cost needs to be reduced. That people need to be let go. It’s easy to look at a credit rating score and use that to turn down a loan application, but does this give the full story?

We find that within any of our clients, the data may well tell a story. We can create models that predict whether someone will behave a particular way, much like meteorologists predict whether it’s going to rain. But the emotional investment of a raindrop is not the same as the emotional investment in someone’s life.

Recently my friend Meagan Longoria, an expert on data visualisation, wrote at https://datasavvy.me/2020/03/26/stress-cases-and-data-visualization/ about how we need to remember that the people looking at data need to understand the stress level that consumers of data are under. The current situation with COVID-19 a strong example she gives about how everyone has been impacted in some way by this virus, whether they’ve been in hospital from it or lost loved ones through to being inconvenienced by closure of their local gym. Some people might happily look at charts about the number of cases all day, while other might be petrified in fear.

I’m not about to start producing charts in the public domain about a subject area I know relatively little about. Particularly when those data points are people who are hurting, and all the data points that are not on there are people who are potentially fearing for their own lives. If a customer of ours wants me to do some analysis on their COVID-19 data, then that’s up to them, and might help them understand various aspects. But the human element of it is not lost there, because it’s an organisation trying to understand the data in their world.

Pay attention to what Meagan and others are saying, and don’t be tempted to report on data just because it’s there. Consider the human impact of it. Seek to understand how people are impacted by the data. Seek to understand what aspects of people are described by data points, by trends in the data, by your reporting of the data.

People are always more important than data, and we would do well not to forget that.

In summary, we need to maintain an empathetic perspective on the data we’re analysing, just like we need to keep an empathetic perspective on people.

@rob_farley