5 Easy Ways to Find Great Stories In Your Data
By Katy French
Data storytelling is one of the best tools out there for content marketers. But for data noobs it can seem super intimidating. Where do you get data? What do you do when you have it? How do you find stories in data? Relax. We’ve been doing this a while, and we’re here to help you get through it.
How to Find Stories In Data
Good stories don’t just come from data; they’re actually hidden in data relationships. When you start to play with your data, you begin to see how each data point relates to another. The patterns you see (or don’t) help uncover what—if any—story is there. Understanding what type of data relationships to look for helps you find those stories faster. But first, let’s guide you through the steps to get to that point.
Step 1: Get Your Data
This is where most marketers get tripped up. You have a spreadsheet in front of you with a few or a million data points. The first step? Make sure it’s clean and organized.
Organize your data: Most of the time you’ll be working with data from a spreadsheet. The format of your data depends on what kind you have. Let’s talk about different kinds of data.
- Is this data one point in time? For example, If you have data from a 2017 survey, you’d have survey questions in the column and answers in the rows.
- Are there multiple time periods with only one observation? For example, if you have data on Apple stock prices from 1990-2016, the format would have years in the rows and the variable or stock prices in the columns. Note: If years and the variable are switched, no big deal. Spreadsheets have a function where you can paste the values “Transposed.” This will switch the rows and columns of the data.
- What if your data has multiple observations over a time period? Let’s say you have a dataset that has info on multiple countries from 1990-2016. This data will still have years in the rows, but each column will specify which observation is for that particular year. In this example, you would have a “country” variable that identifies which country the data is referring to.
Identify missing values or bad data: These make you a less credible source since your statistics will be wrong. Do a visual inspection to make sure that the data points make sense. For example, if the data set measures human weights, does it make sense for someone to be 2,000 pounds? Get rid of rows where there are tons of missing data.
Look for outliers in your data: These would be data points that don’t seem to fall into your range of expectations. Outliers are usually thought of as a nuisance, but they could also offer interesting stories and insights. For example, if we expect sales to go down in all counties, then a spike in sales in one county would be an outlier (more on that later).
Step 2: Visualize Your Data
When we talk about data visualization Go to the full article.