Data best practices vlog with BI specialist, Joseph Hobbs. Episode 1 discusses the common mistake of collecting data for the sake of data.

Video Transcription

Hello everyone, my name is Joseph Hobbs. I’m a Digital Insights Consultant at Valorem Reply and I’m happy to introduce our video blog series ‘This Not That’! The purpose of ‘This Not That’ (TNT) is to contrast a best practice, the ‘THIS’, with a very common mistake that we see as we interact with all of our clients, the ‘THAT’, and encourage you ‘do this thing’ not ‘that thing’.

We’re designing each episode of our series to cover very simple, easy to apply concepts that fall into one of the 5 stages of business intelligence:

1. Get data. This is the data ingestion stage. If you’ve got an IoT project with devices sending information in and you're storing everything in excel sheets or CSVs, or if you have a SQL server set up and you want to move your data there, whatever your environment is, the first step is always to get the data and bring it all in to one place.

2. Clean the data. This is where you’re going to start talking about writing queries, getting things cleaned up and ready. This is where you take the information and ask yourself:

• What do we actually want to know from the data that we’re getting?
• Does the data tell us that or do we need to perform some calculations in order to get the information out of it that is necessary?

3. Analysis. This is where you want to start talking to data scientists and data visualizers who are going to come in and ask:

• What insight do we want out of this information that we have cleaned?
• Now that we have the tools and raw building materials, what do we want to build?
• What is it that we’re going to make at the end of the day?
• How will it impact our business?

4. Visualization. The human brain is very adept at seeing things but only mediocre at reading things. So visualizing data using modern tools like Power BI, enables you to take what would be a convoluted mess of numbers and pieces of information and text and present it in a visual way to more clearly and quickly see the outliers. 

5. Solicit feedback. When you take everything that you’ve ingested, cleaned, analyzed and visualized and you take it to the end users and get their feedback. What do they think about what you've collected and how you are analyzing and presenting it? Does it really solve their problem(s), or has it missed the mark? This is also where you’re going to do your training, creating tutorial videos, doing one-on-one sessions with your end users to make sure they understand, adopt and properly use the tools that you’ve made.

Each of our best practices will fit into one of those 5 categories, some of them will be tool specific, some will be about SQL or Tableau, others will more generic things where we talk about things like Gestalt design principles, which have a broad range of applications.

Episode 1: Insight not Info

For our very first episode today, we are going to be talking about one of the more common requests that I am regularly asked for, information. There’s a lot of hype around Big Data. Everyone knows that the top 5 companies in the world are technology companies- Google, Microsoft, Facebook- who are harvesting data and turning it into gold. So everyone knows they need to go out there and get this Big Data because of all of this hype and build up, but what I want to remind people of- the best practice vs. the common mistake-  is that a common mistake is to focus on the information instead of the INSIGHT that that information provides.

So what do I mean by that? Visualize a raw Excel sheet. Take off the headers at the top and just imagine that you have a bunch of columns of miscellaneous data: numbers, words, percentages. What does it all mean? Well, it doesn’t mean anything, it’s absolutely meaningless without headers to provide context around the data you're seeing. But when you start putting some context around it with column headers, you can easily identify the data you're seeing like a zip code and a percentage of how close you were to hitting your targets. The headers allow you to turn a spreadsheet of random information into actual insight, which is what you you were trying to do in the first place.

So as a rule, if you’re looking at a project, yes the information is necessary, yes we need to focus on the info, but it is NEVER the end goal. The end goal is always to say what insight will the end users get that will impact their actions.

Thank you for following us here on our first episode of TNT. If you enjoyed this, we’d love to see some comments below. If you have an idea for a data best practice that you'd like to discuss, also leave us a comment. Otherwise follow us on social media or head to our website for more episodes of This Not That.