Python, not Tableau, should be your next dalliance with data analysis outside of Excel. When I talk to people about data, invariably I hear the same refrain about Tableau. Everyone believes it is a silver bullet that will solve their heretofore Excel-driven data challenges. I hate to break it to you, but Tableau is not your answer.
Tableau allows users to make some fantastic charts. I’m not some type of Excel Ebenezer Scrooge - I can appreciate an attractive graphic. However, I think if you look within your soul you’ll see that a tree map is not going to unlock the mysteries of your data.
You might be willing to admit that it is pretty hard to build a cool graphic before you understand what you are trying to present to end users. But as analysts, most of us have spent a ton of time working with our data looking for nuggets of insight to share. Our visualization, whether it be in Tableau or Excel, serves an explanatory data analysis function. When building visualizations, most of us are looking for more attractive and engaging ways to explain the insight we’ve gleaned from our data.
Unfortunately for you, what Tableau is best suited for is exploratory data analysis. This is when you don’t know much about your data, and you’re looking to quickly get a sense of what it looks like. If you have data that is coming in fast or in a variety of different schemas, then this step, exploratory data analysis, might be something you need to regularly perform in your analytic workflow. But if you know your data well, then it probably isn’t the biggest bang for your buck.
I’m assuming that you understand the data you are analyzing quite well. You are drawn to Tableau because you want to put a slick visualization in front your boss. I understand this desire - it makes perfect sense. In the analytic world, our carpentry tools are usually Excel and our edifices are charts placed in PowerPoint decks. Not only do your charts look like everyone else’s, but you are just bored of creating the same bar charts over and over.
If you have your heart set on building visualizations, it would behoove your wallet to check out some free charting options, some of which have far more power than Tableau (warning: you will learn to write some code). But if you really want to see your boss’s mind blown, I suggest you look at analyzing your data rather than visualizing it.
I’m sure you have heard that 90% of a data analysts time is spent cleaning data. How does Tableau help you clean your data? It doesn’t. It does have database connections, but so does Excel (and let’s be honest: how often do you use those?). It can ingest JSON and XML, but how does it help you parse it? And even if you succeed in extracting just the fields you want, how do you combine your JSON data with the data from your relational database?
I’m not here to tell you that gathering data from multiple sources and standardizing it for your analysis is easy. But with Python, it is possible. With Python as my helper, I shall fear no data...