The risk of leaning too much on data

- [Instructor] Now let's talk about the problem of being data-driven. If data is the only thing you're using to drive your decision making process, I'd argue that your process is just a little bit too rigid, and it can negatively affect your ability to solve business problems. So now let's walk through an analogy. Within statistical modeling, there's this concept called overfitting a model, and this is where your model is vulnerable to random errors in the data, and that can lead you to the wrong conclusion. This is very similar to if you're using only data to drive your decisions, you are essentially overfitting your model. You may run across data quality errors or mistakes within the data. So what might be influencing your data that can lead you to the wrong conclusion? The first thing is data quality errors. If there is a stake in your database and you're only leaning on the data from your database to inform your decision, then you might come to the wrong outcome. Next we have miscalculations. If your analysis is done by a human, then there's a chance that a human might make an error, and this can also lead you to the wrong outcome. And finally, we have unquantified variables. It's almost impossible to identify every single variable that has an effect or impact on a specific data point. So if you're not taken into consideration all of the variables, then you may be led astray. So to just remember, don't lean too heavily on your data to make every single decision.