Today at work I was asked how I approach a data set I am unfamiliar with, and honestly 8 months ago I’d have a very different answer than I do today.
#MakeOverMonday has really helps my approach to data analysis and much as it has raised my level of Tableau output. Each week, getting a new data set with no fences is a challenge when you first start out.
So, with that said, here are the steps I’ll run through when I get a new data set.
First; Play with it! Nothings is as exciting as Sheet 1 in a new workbook, because at that point your free to go whatever direction you want. Drag on fields, sort them, filter them, use the Show Me tool, and explore the data. Sometime, the data is easy and patterns and stories start to surface as you twist and turn the data. This is especially true in smaller data sets.
Second; Make it personal! What’s important to you? Is there anything in the data you can tie directly to your life? This is a great technique in large data sets. Today, we were looking at ATM utilization data at work, and as you can image, that dataset is HUGE. More than a billion transaction records in the fact table, and ATMs all around the country. How do you filter that down? I made it personal, I picked the small town I live in, and it filters out more than 90% of the data. Using that subset, we can start to answer some questions, then apply it to the larger data set to see if the same assumption apply.
In the case of Makeover Monday the personal story is enough to present. Several times I’ve used geography from my life to find the stories, whether it was a place I lived, or my family’s origins and homeland.
Third; When in doubt, Google it! Sometimes the news can help inspire the story, so a quick google of the subject are of the data and help guide the analysis.
If this doesn’t help, and you are still struggling with the analysis and story-finding, then I recommend finding some data that you find interesting! Practice really does help in this area, and the more you do it, the more comfortable you will become. Love Dave Matthews Band? Go to wikipedia and throw his discography and it’s performance in Excel. Maybe find their touring history and map it out? Wikipedia has tons of data you can scrap off and use for practice.
There are also a bunch of resource for data sets, some are better then others but there are out there. Here are a couple of good ones I’ve found:
Now, get out there are find those data stories!