Short thread: The greatest challenge facing data analysis

We often debate about viz best practices. If we need Dashboards, alerts, or presentations. If we should call data presentations stories. Are infographics good/bad?

Important, but misses the key challenge

1/n

We have a lot of hubris in assuming showing a bar over a pie chart will immediately change how people view data. Yes, it will certainly help visualize the insight. But...

Pie vs Bar chart, Dashboard vs Alert doesn’t matter if the audience doesn’t trust or care about data

2/n

We have to realize many people are hesitant to trust data. They are willing to make decisions against data visualizations/stats because for a variety of reasons. Or trust bad data if it fits a pre-conceived view.

This isn’t an edge case. This is more common than we think

3/n

We can clearly show anomaly and visualize correlation vs causation, but we can’t use our “best practices” or “UX” as a crux. We can’t build charts and think our job is done.

We have figure, somehow, to convince our audience to trust data, and to trust us.

4/n

Our role as analysts isn’t just to analyze data and make charts. Our role is more important than that. We have to convince people to trust us, our methods, and show them we are trying to help answer questions based on facts.

It’s a lot to ask of a person.

5/n

It’s part of the job. To be more than just a chart maker.

Sometimes data is clear to those who use data to make decisions, but not to those who don’t or are driven by emotion.

We have to get them to trust the process, before they can trust the data

6/n

If users don’t trust you, don’t trust the process, or don’t trust the argument, they will never trust the data. It doesn’t matter about the visuals used, dashboards, UX, or anything else.

We should about this more.

End random rant.