[MUSIC] Welcome back. The human brain loves stories, but this is a double edged sword. Sometimes we see a story where there isn't one. As a data storyteller, you have a lot of potential power and responsibility. How you select, arrange and present the facts can change the way people think or how they view the world. You should strive to make your data stories reflect reality, however, even if you have the best intentions that's not always enough. In the process of assembling your data stories, we have to be diligent to avoid creating false narratives. You need to check your own unconscious cognitive biases and consider potential biases in your audience. That's what we'll be examining next. After this lesson, you will be able to recognize cognitive biases and enact counter-measures to guard against them. You will also be able to take steps to avoid presenting false narratives. Let's get started. Hi, and welcome back. Data visualizations reinforce our innate tendency to find patterns and convert them into stories. If we see what appears to be a close correlation between two things, there can be an almost automatic attempt to find or create a story to explain that apparent connection. For example, look at these charts, which I'm partially reviewing, so you can focus on the lines. There seem to be some pretty close relationships, and some interesting details like the inflection points in the second chart. What's the story within that? Well, nothing really. These examples come from the amusing site, Spurious Correlations. Here's a full view of that first chart. The relationship being displayed is between worldwide non-commercial space launches and awarded sociology doctorates. And here's the second chart. We see a pattern of the two things being measured. In this case, the age of Miss America and murders by steam, hot vapors and hot objects. They seem to track to each other very closely, and then we have what might be perceived as interesting inflection points, all of which are ultimately meaningless. What makes these ridiculous combinations so great is precisely that, well, they're so ridiculous. We don't need to be subject matter experts to know intuitively that these relationships are false. However, there are many instances where it may not be so obvious at all whether a correlation is meaningful. Visualizations can make these correlations seem more intriguing and tangible, and so the patterns get turned into false narratives. This may be an even greater danger and challenge if you are not a subject matter expert in the domain from which your story is being created. I found myself in that situation, as perhaps some of you will as well. Part of the process of guarding against false narratives involves talking to subject matter experts or SMEs as we touched on in Module One of this course. It's also a good idea to do some outside research whenever possible as well as consider other data sources to help you verify that you're not unintentionally creating fiction. It's all too easy to impose a story on a disconnected set of facts that appear to relate to each other. These are sometimes called, just so stories, or in some contexts, ad hoc fallacies. The term, just so stories, comes from the title of a children's book by the author, Rudyard Kipling, with whimsical tales of how some animals developed certain physical attributes. But these kinds of observations can be quite useful, and it can sometimes be used as hypotheses for further testing and investigation. However, stories that oppose an unverifiable explanation can be very problematic. The bottom line is that people can create all kinds of origin or causative tales about patterns that they see. But huge expansion of data to analyse and visualizations like Tableau to work with data, increases the opportunity for creating both true and false data stories. Now, along with the tendency to create false narratives from unconnected patterns and data, there are a number of ways that data storage can be unintentionally twisted by their creator's cognitive biases. Basically, cognitive biases are tendencies to think and make judgements based on our own personal set of perspectives, experiences and filters. And that may not be reflective of the true situation. In general, there are a number of counter measures against forming cognitive biases. The first step is cultivating an awareness of the potential for cognitive bias in yourself and others. After that, for any data story that you create, be sure to consider other possibilities. Even stories that seem diametrically opposed to your own ideas and conclusions. Additionally, whenever possible, it can be very helpful to test early versions of your data stories on people who are in, or are similar to, your target audiences. You may find that the stories evoke common responses that may not have been your original intention or anticipation. If possible, it can also be a useful exercise to test out the story in a variety of people outside of the target audience. There are many potential biases and logical errors that most of us can fall into. The following info graphic catalogs and categorizes quite a few of them. Now, do I think you need to memorize all of these to become a good data story teller? No, however, it is important to develop an appreciation for the many ways that we can fool ourselves and others with false stories. That way you can be alert throughout your story development process. Why am I putting so much emphasis on bias in this lesson? Well, data stories are the place where either truth or fiction can be vividly brought to life and disseminated to many others. I hope all of your data stories are engaging and accurate ones. Well, this wraps up our view of being alert to the dangers of creating false narratives. In our next lesson, we will revisit the tale of 100 entrepreneurs, and walk through it to learn how to tell a specific story from a data set that contains many potential stories. Thanks, and see you again soon.
