Data assimilation interpretation

A Non-Analyst’s Guide to Data Assimilation and Interpretation

Unlike data analysis, the goal of data assimilation and interpretation is to (1) get oriented to the information in front of you and (2) form an initial impression. This post is designed to help people who are not full-time analysts properly assimilate and interpret the data in front of them each and every time.

The Three Common Flavors of Non-Analysts

In my experience, non-analysts often move way too fast when they encounter data. Anecdotally, I find these consumers of data fall into three main categories:

The Stakeholder

This group is usually hungry for information. They are often project leaders who have a genuine stake in the results of an analysis or the data in general. 

These data consumers are typically so busy with their day-to-day work, that when they finally have a chance to sit down and look at the data, they’re already fatigued or preoccupied. They also are the people most likely to “see what they want to see” because of the anxiety they have around the results. 

You see where this is going.

The Unequipped and Disconnected Audience

I can’t tell you how often I’ve sat in an “All Staff Meeting” (or similar gathering) while a presenter bombarded me with information for which I had zero context. It usually soon becomes apparent that I’m not alone in my disorientation. There is generally a whole group of us in the Unequipped and Disconnected Audience.

People in this group are out of the loop. Team leads and management have not adequately equipped or prepared then to receive the information. It just gets dropped on them. To catch-up, this group hastily tries to put the pieces together, and they often misinterpret information miss invaluable details.

The Non-Analyst Data Enthusiast

They may not be full-time data professionals, but these people go out of their way to learn data skills. They then frequently use those skills to actively incorporate data into their strategies and processes. Unfortunately, they often also have significant gaps in their knowledge of analytics. 

I’m not saying that everyone needs formal training to become good analysts. Having ample expertise working in other roles within a field is an advantage in analytics work! Yet, unless they’re devoted full time to analysis, they’re probably going to miss some things (or do some things incorrectly).

How to Assimilate and Interpret Data Correctly (in Three Acts)

Each of the groups mentioned above could mitigate much of their data assimilation and interpretation issues by following some systematic workflows.

Workflow 1: The Stakeholder

This group struggles with cognitive bias, distraction, and anxiety. Their workflow emphasizes objectivity and tries to minimize gut-reactions. Since this group is the group most likely to consume data after it’s been analyzed by others, this workflow also emphasizes careful comprehension.

Schedule dedicated time for review. You need to set aside dedicated time for this. Grab some coffee, go to your most productive and distraction-free space, turn off the email notifications, and relax. You may also want to grab a notebook for jotting down thoughts as you go. You need to be focused when you assimilate and interpret data.

Identify the topic of the data and the source. Read all titles, subtitles, and descriptions. Determine if the source is legitimate and trustworthy. If you don’t recognize the source, look it up. Only move forward once you’re convinced the source is reliable.

Get familiar with the axis and units of measurement. Be sure you’re aware of factors like scale and obscure calculated metrics. Also, check for structures like dual y-axis that can make comparisons between data difficult. If the data is in the form of tables, be sure you’re aware of how the values in each cell were calculated.

Verify that the data actually represents what you think it represents. Sometimes data has been filtered before being shared. This is particularly prevalent in shared environments where multiple parties have edit rights. Regardless of the conditions, be sure you’re getting the view you think you’re getting.

Start observing patterns in the data. Without passing judgment, look for common patterns like skewness, clustering, and trends. Also, look for any outliers and noise (extraneous information) in the data.

Be sure you’re comfortable with your first-pass comprehension. Do you have a clear picture? Are you satisfied that the data is presented in a way that will allow you to form some preliminary conclusions? Do you need to go back to the analysts and ask for some clarifying information?

Form your preliminary conclusions. Once you’re comfortable, it’s time to form your initial conclusions. Make some notations that describe what you’re seeing. Record any questions you have about the information. These notations should be free form and will inform your next steps.

Enlist a second pair of eyes. Since you are inherently prone to cognitive bias as a stakeholder, enlist a knowledgeable (non-stakeholder) peer to give you a fresh, neutral pair of eyes. See if their interpretation aligns with yours. Avoid looking at data for the first time in a group of fellow stakeholders. Factors like peer-pressure and groupthink can cloud your judgment.

Establish your next steps, and a means to answer any follow-up questions. Assuming you and your neutral second pair of eyes are aligned, you should outline the next steps that need to be taken. If the data raised more questions, establish a plan to get the answers. If you’re confident you have the conclusions you need, then those conclusions should be shared. Don’t let data silo and stagnate.

Workflow 2: The Unequipped and Disconnected Audience

Recall that this group struggles with the “surprise factor.” This workflow attempts to provide a re-orientation pathway. It re-aligns teams to ensure everyone is getting the same picture. Like the Stakeholder group, this group is also likely to be consuming the analysis of others.

Pull the emergency brake if necessary. Hopefully, you have a good enough relationship with the data presenters that you can to ask them to pause. If you and your peers need more processing time, say something! Chances are that the person presenting the data will want you to get the most out of the information and will happily give everyone a moment to digest.

Be sure you have context. You cannot assimilate and interpret data effectively without context. If you’re hearing about this information for the first time, then you should say something. Ask clarifying questions. Be sure you are prepared to go forward.

Get grounded and establish the data source. Read all titles, subtitles, and descriptions. Determine if the source is legitimate and trustworthy. If you don’t recognize the source, look it up. Only move forward once you’re convinced the source is reliable.

Get familiar with the axis and units of measurement. Be sure you’re aware of factors like scale and obscure calculated metrics. Also, check for structures like dual y-axis that can make comparisons between data difficult. If the data is in the form of tables, be sure you’re aware of how the values in each cell were calculated.

Verify that the data actually represents what it claims it represents. Be sure the data is used appropriately and that you’re getting the view you think you’re getting. If you’re interacting with the data in a shared environment, be sure that you’re aware of any filters or sorting that may be active.

Start observing patterns in the data. Without passing judgment, look for common patterns like skewness, clustering, and trends. Also, look for any outliers and noise (extraneous information) in the data.

Form your initial conclusion and check for alignment with the presentation. Did you assimilate and interpret the data in the same way as your peers? How about the same way as the presenter? Do you think that the presenter may be presenting a biased view of the information? Does the data actually support any conclusions you’ve been told? Push back if you feel you’re fed inaccurate information!

Be sure you’re comfortable with your first-pass comprehension. Do you have a clear picture? Are you satisfied that the data is presented in a way that will allow you to form some preliminary conclusions? Do you need to go back to the analysts and ask for some clarifying information?

Have discussions with peers if something isn’t clear. If you’re unsure about anything in the data or the analysis, get input from your peers. Constructive group investigation can be a powerful data assimilate and interpretation technique.

Provide feedback to improve future presentations of information. If the presenter provides a feedback mechanism, be sure to ask for ample time to assimilate and interpret any data. A good data presenter will always bake-in plenty of time for the audience to digest the information in front of them. The best data presenters also provide ample time for discussion. Also, ask your team leads to provide sufficient context in advance of future presentations.

Workflow 3: The Non-Analyst Data Enthusiast

Even though this group is more equipped than their non-data-inclined peers, they still need to be careful. Their workflow attempts to ensure that they maintain a balance between their mainline work responsibilities and their analytic enthusiasm. It also strives to keep them grounded, methodical, and less prone to making errors due to knowledge gaps.

Get in the right headspace and move slowly. Set aside plenty of time for review. If you’re in this group, you’re naturally inquisitive. Be sure you have time to chase all the questions you’re going to have. Also, be sure that you don’t have higher-priority things to do that should be handled first. You want this experience to be stress-free.

Be sure you have context. If this data is coming out of the blue, take the time to establish context. Don’t make assumptions or jump into this blind.

Identify the topic of the data and the source. Read all titles, subtitles, and descriptions. Determine if the source is legitimate and trustworthy. Only move forward if you’re convinced the source is reliable.

Get familiar with the axis and units of measurement. Be sure you’re aware of factors like scale and obscure calculated metrics. Also, check for structures like dual y-axis that can confound data comparisons. If the data is in the form of tables, be sure you’re aware of how the values in each cell were calculated.

Verify that the data actually represents what it claims to represent. Sometimes data has been filtered before being shared. This is particularly prevalent in shared environments where multiple parties have edit rights. Be sure you’re getting the view you think you’re getting.

Start observing patterns in the data. Without passing judgment, look for common patterns like skewness, clustering, and trends. Also, look for any outliers and noise (extraneous information) in the data.

Test any statistical assumptions in the data. This is really much more analysis than interpretation, but it’s something you should think about. Unlike your more non-data inclined peers, you’re very critical of things like sample sizes and skewed averages. Pause here and take plenty of time to put the data through the paces and ensure any assumptions are valid.

Be sure you’re comfortable with your first-pass comprehension. Do you have a clear picture? Are you satisfied that the data is presented in a way that will allow you to form some preliminary conclusions? Do you need to go back to the analysts and ask for some clarifying information?

Form your preliminary conclusions. If you’re comfortable with your data assimilation and interpretation up this point, it’s time to develop your initial judgment.

Run your conclusions by an analyst peer. You don’t know what you don’t know. It’s often a good idea to get a trained pair of eyes on your conclusions before going forward. Even full-time data pros rely on peer input to form complete, accurate pictures. There is absolutely NO SHAME in collaboration. Peer-review never diminishes the validity of your work. If anything, it enhances it!

Establish your next steps. When you’re confident with your conclusions, outline any next steps that need to be taken. If the data raised more questions than presented answers, establish a plan to get those answers. If you’re confident you have the conclusions you need, then those conclusions should be shared. Don’t let data silo and stagnate!

Continuously Develop Your Ability to Assimilate and Interpret Data

Regardless of what group you fall into, learning to assimilate and interpret data is a continuous process. The range of analytic techniques and data sources is always expanding. You’ll need to keep your skills sharp and up-to-date at all times. There is no such thing as peak data literacy. Also, be sure you make those skills regular parts of your work. If you don’t use them frequently, you risk having them stagnate. It’s not a skillset, it’s a lifestyle 🤓

Advertisements Disclosure

I will always make it clear if I am writing to endorse or recommend a specific product(s) or service(s). I hate it when I visit a site only to find out that the article is just one big ad.

Various ads may be displayed on this post to help defray the operating cost of this blog. I may make a small commission on any purchases you make by clicking on those advertisements. Thank you for supporting my work bringing you accurate and actionable information on data literacy, analytics, and engineering.

Advertisements

One Response

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.