Does it feel like data is the road block to progress in your organization? Where is the data? What data are we waiting on? Do we have better data? These questions and frustrations are more than common. They can feel all-consuming when your time is spent waiting on data requests to be fulfilled by a few busy analysts with the only alternative being long, involved software trainings that still don’t leave you prepared to get what you need.
What if you didn’t have to wait any longer? What if you had the power to answer your data questions instantly and go deeper? Metopio is a platform for finding and using data about places across the country, including your community. The platform’s curated datasets and easy-to-use analytics tools make it easy to find information, explore it, apply it, and share it.
We know that this may sound too good to be true. It’s not! With a new account you can download datasets and visualizations about your community in under two minutes. We know that data can feel overwhelming. We also know that the scariest part of diving into data can be the fear of making mistakes, which can turn your analysis from accurate to erroneous.
We’ve highlighted 3 data mistakes to get you started –why they matter, and most importantly, what to do instead:
Mistake #1: Not comparing apples to apples
In Metopio, insights aren’t about individuals and data is tied to a place. So we don’t track whether Sonya Smith has a high school degree. Instead, we assign data to the ZIP code and then provide the graduation rate for that zip code. Not tracking individuals protects privacy. It also enables us to mix and match data sets. Our data model is based on places which are all defined the same way. The same ZIP code in one data set is the same ZIP code in another. This method of data curation allows you to use topics interchangeably and share data and your insights without risk.
However, there is also count data in the platform. When you’re using data that is a count of something (like “number of customers” or “total widgets sold”), it’s important to be aware of comparing unequal populations. Of course a ZIP code with more people has more customers – that makes sense! While rates and percentages allow you to compare unequal populations, count data does not.
When does count data come into play? If you’re looking to understand volumes, but not comparisons, count data can be used effectively on a heat map. For instance, “Where do most of my customers come from?” as opposed to “Where are people more likely to be my customer?” Otherwise, stick with rates to stay safe!
Not comparing apples to apples can also be a problem when choosing benchmarks. Benchmarks, when used wisely, are one of best ways to include context to help explain the story behind your data.
Let’s say you’re looking at public transportation to work in the city of Chicago. It would make sense for you to add benchmarks that represent similar or nested units. You may then compare Chicago with another large city like Aurora, Illinois. Alternatively, you could also compare the use of public transportation in Chicago with a ZIP Code like 60641, Cook County and Illinois to see how you measure up. But comparing Chicago to rural Illinois probably does not make sense, because the environment, lifestyle, access to transportation, and people are very different.
Mistake #2: Missing the margin of error – and missing its meaning
When dealing in data, the margin of error is a metric that you must know – and keep top-of-mind. It allows you to gauge the level of unpredictability in data.
As a rule of thumb, the smaller the data set, the larger the margin of error. Paying close attention to the margin of error is especially critical when using small data sets because the data has a less accurate depiction of the entire population.
TIP: look to see if the margins of error on two data points overlap and if so, it can be dangerous and inaccurate to say that one data point is definitely higher or lower than the other. Another red flag is when you see very different values from one year to the next like 50 to 3000 to 2. This is also likely a small sample size issue. Using small data sets is only meaningful after accounting for the margin of error.
What is the “danger zone” for a margin of error? Having a margin of error that is more than half of the total size of the statistic is too uncertain. It might be better to remove or suppress (aka ignore) that data.
Mistake #3: Skipping the story behind the analysis
The goal of any data analysis should be to tell a compelling story.
Data points need context to tell a story. They rarely, if ever, tell a story on their own. You can answer a single question but you need to build confidence and certainty to round out your narrative. If you did an analysis of where rents are the highest, you must also account for the average income of individuals in that place. High rents may not be equitable, but they may be attainable if the average income of individuals in the community support it.
That’s why we suggest never approaching a single dataset on its own. Ask questions! You can’t break Metopio. Think about why you’re asking the question. Unpacking your question enables you to understand what other questions you should ask and answer to explain your findings.
By digging deeper, you can discover what else you, or your audience, need to know to make a smart decision about the topic. Ensuring that all of your insights have a title and caption also helps to provide much-needed context.
Data at Metopio
Metopio continually updates and adds curated, verified data and provides ridiculously easy tools to understand populations and places you care about. All you need to do is point, click and search to build the insights you want. Then, share your insights — without having the mind of statistical savant or programming whiz – and without having to wrangle the data yourself.
Is our mission aligned with your goals to do more with data? Schedule time to speak with our experts here to discuss the questions you are asking today.