Poor workplace data literacy is pervasive. While things like data rot and data stagnation will wreak havoc on your operations, not establishing and improving data literacy at your organization will completely obstruct all your data-driven dreams.

Before you can reap the benefits of all your hard-collected data, you must establish a comprehensive data literacy plan. Trust me, this work is well worth the investment, and the benefits are immediate.

Data-Driven (Right into the Ground)

Let’s set the stage. Your organization is a “data-driven workplace” You’ve purchased all the right data tools. You’ve consulted all the right consultants. Your capable data analysts, engineers, and scientists are building some amazing things. You have clarity around your information like never before. 

Despite all that, however, your organization’s growth and productivity have stalled. The data resources that were supposed to be an elixir don’t seem to be doing much. What gives?

At this point, organizations usually go down one of three ineffective paths in an attempt to fix the problem:

Data Failure Path 1 | Abandon and Re-Do

The first path is a progressive abandonment of existing resources. Soon after, there is a flurry of vendor bids and another familiar cycle of “exciting new tools and improvements.” New toys are purchased, new processes implemented, and the nightmare begins anew.

This path is usually the gut-reaction, shoot-from-the-hip management response, and it is as woefully ineffective as it is hasty. “Turning and burning” away from the root problems will get you nowhere.

Data Failure Path 2 | Data Rot and Systems Degradation

Whereas the first path is sudden and in-your-face, the second path is insidious. Existing tools gradually degrade or fall out of use. Clean data loses its integrity. Ultimately, confidence erodes entirely. 

Teams cope with this vacuum by creating a disjoint network of resources. If you’ve been through this scenario, you’re familiar with the patchwork of spreadsheets, forms, and redundant processes that can result. 

Eventually, teams are spending so much time duct-taping things together, that they don’t have any capacity to move the organization forward.

Data Failure Path 3 | Incomplete Diagnosis and Repair

This path is the most enlightened of the three fallout scenarios, but it still falls short if the issue is poor data literacy. On this trajectory, team leads re-evaluate their processes and execution. They collect feedback and see what went wrong. After a thorough investigation and many, many meetings, they lay out a repair plan to get the organization back on track.

Unfortunately, such efforts often fall short, and the team finds itself continuing to struggle. Why? The plan might treat the surface symptoms, but it likely neglects an underlying disease: poor data literacy.

Workplace Data Literacy and Why it Matters

Data literacy means what it sounds like it means: the ability to appropriately interpret data and communicate findings effectively. Just like other forms of literacy, you have to learn this ability. Yet, for some reason, team leaders often expect that it just manifests out of thin air. You cannot expect your team members to possess this literacy automatically.

Data literacy means what it sounds like it means: the ability to appropriately interpret data and communicate findings effectively. Just like other forms of literacy, you have to learn this ability.

The consequences of not fostering and improving data literacy at your organization can be crippling. Invaluable insights are easily missed or misinterpreted. Assumptions and conjecture frequently replace facts or best-practices. What should be an enthusiastic environment of curiosity and critical thinking devolves into frustration and apathy. 

Eventually, day to day life at the organization becomes a stressful mess, and everyone burns-out.

Key Considerations for Improving Data Literacy

Improving data literacy is a multi-part process. How you put the pieces together will depend on the unique conditions of your work environment. You should, however, keep some universal core principles in mind:

Make your sessions adaptable. Structure your learning sessions so that they can be rearranged (or outright changed) based on the response from your students. You may need to throw the orImprovingiginal game plan out the window if it looks like your students require more/less time with a particular data literacy concept.

Keep your sessions relevant to the students’ present work. If you can make it clear how your topics directly link to the problems your students have now, they will learn much more quickly and effectively.

Keep sessions in-line with students’ personal goals. It should be immediately apparent how what they learn will empower where they want to go.

Be fairly casual. Your learning sessions should not be lectures. Set aside ample space in each session for collaboration and discussion to help crystallize the material. The key here is to make them active and welcoming. Just don’t make it theater or fill the time with pointless “icebreakers” and other ineffective pseudo-team-building fillers.

Still with me? Good! Let’s move onto the steps you can take to improve workplace data literacy.

Get Prepared

Understand the amount of time, mental energy, and resources this will take

Your organization’s investment in improving data literacy has to be top-down, and it has to be real. You cannot half-commit to this process. Everyone must be on-board.

To that end, you’ll need to do substantial work before you can even begin to tackle things at the skills level. Be sure you’re mentally prepared to take on this work, or you’ll burn-out.

Build a small team of data literacy project leads

You should not undertake this initiative on your own. You probably already have a shortlist of data allies in mind, so consider if any of them would make good co-leaders for this project. 

If you’re an introvert, this is particularly necessary. This whole data literacy improvement process is going to require some well-developed people skills, so be sure that your data literacy “Steering Committee” has members who are comfortable communicating to groups and navigating office politics.

You also obviously need to make sure that you have some of your organization’s senior data folks involved at the planning level. They will help ensure you are covering the right topics and that your information is correct. They’re also the ones with the best perspective of your organization’s data processes in situ.

Crystallize your vision of workplace data literacy before you get tactical

It’s critical that you (and your Data Literacy Steering Committee) have a clear sense of purpose before you get into the details. You should be able to quickly and effectively address the following if asked:

  • Why are we embarking on this initiative?
  • How will this initiative improve the well-being and productivity of the organization?
  • Why does the initiative need to start now?
  • What are the obvious pressure points and obstacles to success for this initiative?
  • What are the core indicators (e.g., “metrics,” “signs,” etc.) for success with this initiative?
  • Why should anyone care about this initiative? How will it improve their work and help them reach their goals?

The Data Literacy Steering Committee should know the answers to those core questions like the backs of their hands.

Get genuine buy-in from your organization’s leadership

Improving data literacy requires buy-in from your organization’s leadership. That doesn’t just mean the C-Suite. It ultimately needs to come from the top and all the way down. You’ll need this support to secure both the resources you need to make your initiative happen as well as the authority to enforce participation. Words like “enforcement” may come across as abrasive, but that’s not the intent. The intent is to have a backstop to counter the inevitability of resistance to change.

It’s also worth mentioning that this buy-in can’t be passive. You need to get your organization’s leaders to take the Red Pill and follow you all the way.

Include data literacy resources in your organization’s knowledge base

Hopefully, your organization has some type of knowledge base; a place where your workers go to find information about how to perform their work. If you don’t have one, you need one. 

Set aside some space in that knowledge base for some content related to improving data literacy. It doesn’t matter how good of an instructor you are, not everything you teach is going to stick the first time. Set your team up for success by providing informative, well-structured reference material in a familiar and accessible place.

That content will be particularly accessible if your organization uses a knowledge base platform like Zendesk. That type of platform allows users to structure questions intuitively and quickly locate answers. It will also help you track where gaps might be in their understanding of the material because of the robust user analytics.

Make your organization’s data readily available

This section could alternatively be titled, “free your data!” You need to make sure that actionable data is accessible for your organization to use throughout your training plan.

At my current organization, we accomplished this by rolling out a custom data portal web app strapped to our data warehouse. We also crafted some Tableau reports for things not (yet) in the data portal or for other ad-hoc analysis that needed to be shared. 

Your solution may not need to be that detailed. It might just be a repository on your Google Team Drives where you place files that dynamically update from data sources periodically. Do what works for you and your budget. There’s no shame in a suitable working solution that doesn’t look super shiny.

As long as the data is accessible and in a useable format, you’re good to go. For example, if you want your students to start working with Google Sheets, then make sure your data files can be imported into a Google Sheet with minimal effort.

Building Your Learning Sessions

OK. You’ve done all the rigorous prep work, you have the engagement of management, and it’s finally time to get tactical and have some learning sessions. Where to begin?

First, recognize that you aren’t teaching a group of undergrads just trying to make it through the semester. You are teaching a group of your professional peers. Take care not to be pedagogical or patronizing. Instead, use the collective wealth of their experiences to frame their learning.

Part 1 | Core Workplace Data Knowledge

Your first Improving Data Literacy sessions should revolve around core data knowledge related directly to their work. Collectively, these are the data tools, terms, and practices they should be encountering every day.

If your organization is a school, this might mean discussing how your student information systems, data warehouses, and reporting systems all fit together. It might also mean covering why you collected specific data points (e.g., ADA) and how that information affects funding.

If your organization is a hospital, this might mean discussing the interconnectivity of your data systems and where all the information goes. Draw the line directly from a nurse updating a medical record after administering a medication all the way to better patient outcomes. Prove it is worth their time to chart and document various parts of their care.

This is your opportunity to earn your coworkers’ buy-in through very clearly demonstrating why improving data literacy (and data literacy in general) matters to them right now. Be sure they know the following:

  1. The systems they will encounter
  2. Where the information goes
  3. How data fits into the big picture
  4. Any jargon or other unclear terminology they might encounter and how to use that language appropriately
  5. How knowing all this will improve their experience at work and help them achieve their goals

Part 2 | Data Skepticism and Rhetoric

Data analytics has its roots in maths and sciences. In its most elegant and powerful form, it’s as considerate of rigor and objectivity as it is about elucidation. As such, good data analysts should exercise an objective and skeptical mindset regarding all data they encounter. They should ensure they are satisfied that data is accurate and not misleading before they give it the time of day.

This skill is arguably the most crucial data literacy skill of them all: thinking critically. Really let that sink in. It means that you never allow yourself to get caught up in the excitement of data or your findings, and you always assume something is untrue until you can rigorously prove otherwise. Data analysis is an investigative discipline, and your data literacy sessions at this point should focus on cultivating that kind of skeptical, curious, and logical mindset.

That final point is particularly important. Consider the chart below released by Reuters. It depicts gun-deaths in Florida over time, but the y-axis is inverted. A downward trend in the line corresponds to an upward trend in the number of deaths. That’s inexcusably misleading, but you don’t notice the issue unless you diligently criticize the visualization.

improving data literacy
Terribly misleading data visualization originally published by Reuters. The y-axis is inverted relative to the typical orientation, giving the impression that gun deaths dramatically decreased after Florida enacted a “Stand your Ground” law in 2005. In reality, gun deaths in the state dramatically increased.

Creating a Data Skeptical Mindset

Your students should always at least consider the following whenever they encounter data

Was the data collected properly? For instance, is the sample size statistically adequate? Was it a random sample (if necessary)?

If the analysis purports to be relevant today, do the time scales support that argument? For example, is the data possibly too outdated to be related?

Is the source of the data credible? Is it from an organization or individual that may have a conflict of interest with the findings?

Are there comparable studies available to see if the data is in-line with established knowledge? For example, if a dataset seems to show something contrary to other data on the topic, could it just be a mistake? Conversely, is there something in the deviant dataset that might indicate a legitimate shift in established knowledge, and how would that be demonstrated?

Are any findings statistically significant? If not, could they still have practical significance?

What impact could the data have on specific stakeholders, and how will I protect its integrity if those stakeholders try to manipulate the findings?

Has the data been manipulated to mislead the viewer?

Lastly, a good data analyst should have a decent command of rhetoric. They need to be able to articulate their analysis and defend their conclusions and point of view if they’re going to be active agents of insight.

Part 3 | Data Assimilation

Once your data literacy students have a decent working knowledge of your workplace data environment, it’s time to move on to data assimilation.

This stage isn’t about analysis. It is about approaching data, consuming it, and making general observations. Your goal here is to equip your students with strategies that they can use (time and time again) when they encounter data out in the wild. At the bare minimum, this training should include:

An overview of the types of data visualizations they will commonly encounter and the components of those visualizations.

An introduction to common data observations and what they mean (e.g clustering, skewness).

How to objectively approach data and get a sense of its structure without jumping to conclusions or immediately trying to analyze.

Ways that visualization authors can create misleading data visualizations that “spin” facts. This skill is particularly crucial in an era of widespread, coordinated misinformation campaigns.

Formulating a Workflow

Based on the above, a straightforward initial data interpretation flow might look something like this:

Identify the topic of the visualization and the data source.

Determine if the data source is reliable and objective

Determine if there is sufficient information to support subsequent analysis (e.g., are the sample sizes adequate?)

Identify the class of visualization (e.g., line chart) and the components of that visualization

Go oriented to the visualization (e.g., where are the axis and which direction is positive vs. negative)

Identify overt data observation patterns (e.g., clustering)

If you can get your students to follow that kind of flow every time they encounter data, you’ve made tremendous progress toward improving their data literacy. 

Be sure that you’re using examples from their day-to-day work as much as possible to help build upon the great job you did in Part 1. You need to make sure that you’re continually compounding the knowledge layer by layer so that it sticks. Don’t let these sessions devolve into disjoint learning modules. There should be a clear connection between each one.

Part 4 | Basic Descriptive Data Analysis

OK. Your students should now be familiar with their data ecosystem at work, and they’ve learned how to approach data objectively. It’s time to dive into actual data analysis. 

We’re going to stick to fundamental techniques here and focus on basic descriptive analytics. This type of analysis is a hindsight analysis. It studies and describes what already happened.

Before going forward, you’ll need to carefully consider what analytic techniques are most applicable to your particular line of work. You’re looking for straightforward and simple methods that can be readily adopted by your students. You also need to equip them techniques they can use immediately to capture their attention and crystallize their knowledge.

You may also need to spend considerable time equipping them with the tools to do this work. Giving them the knowledge without the means to practice is a one-way ticket to Failure Town. You’ll be the mayor of Failure Town, in fact. So, figure out what they’ll need before you get this point and start getting them primed. 

These tools might be Google Sheets (very accessible, free). They might be something like R (free, awesome power but challenging to learn because of syntax). Or, perhaps you’ve got money to spare, and you put them in Tableau (pricey, powerful, has a learning curve, but that is mitigated by the fact that the platform is attractive.)

But I digress. The point of this stage is to familiarize your students with basic analytic techniques and instruments.

For example, consider the following visualization. It represents quarterly measurements of international tourist nights (in millions) in Australia for several years.

improving data literacy
Quarterly international tourist nights spent in Australia over several years. This plot was generated using the ggplot2 R package.

Based on what they’ve learned in your prior data literacy sessions, your students should feel prepared to summarize the data in a way similar to the following:

The visualization depicts International Tourist Nights Spent in Australia over time. The source of information is the International Visitor Survey conducted by Tourism Research Australia. A quick Google search showed that Tourism Research Australia is an official government entity, so the data comes from a legitimate source.

In this visualization, the total tourist nights (in millions) were aggregated by quarter and plotted as total nights spent in Australia (y-axis) against time (quarters/years). The data is a line chart with each point on the line representing a quarterly measurement.

The data appears to exhibit periodicity and has a steady upward trend.

If they’re able to hit those main points, then they’re ready for descriptive data analysis! 

Data analysis is the beginning of the fun stuff. It’s the chance to dive-in and gain insight from the information. In this example, their analysis might start by checking if the periodicity they noticed in their assimilation/interpretation phase is a seasonal pattern. 

They might accomplish this by first creating seasonal charts like the one below.

improving data literacy
Seasonal plots of quarterly international tourists in Australia over several years. This plot was generated using the forecast R package.

They should notice that there is apparent seasonality in the data with seasonal peaks in Q1, seasonal lows in Q2, and a gradual upward climb through Q3 and Q4.

Your students would probably have more questions at that point and could continue their descriptive analysis from there based on those questions. Be sure they stick to descriptive analysis, however. Don’t dabble in topics like correlation or prediction just yet.

One last bit. Regardless of the examples you use, be sure to use data and scenarios that matter most to your students’ work. You want them using real data from their work so that they instantly recognize the material as relevant. This approach will also encourage them to use the techniques they learn in their daily work.

Part 5 | Correlation Analysis

With a good understanding of how to approach, criticize, consume, interpret, and descriptively analyze data under their belt, your sessions can shift to discussing correlation. 

The goal here is to train your students never to consider data in isolation. Your students should always be on the lookout for different pieces of the same picture.

To that end, have them practice performing simple comparisons of disconnected datasets. They should be encouraged to explore as many possible correlations as they can. Datasets that may seem unrelated on the surface can sometimes have associations that are strong enough to be used for prediction!

R correlation matrix example plot and improving data literacy
It might be useful to demonstrate how multiple correlations can be tested simultaneously in platforms like R. Source.

Be sure, however, to emphasize how easy it is to draw correlations between entirely unrelated data.

Part 6 | Bonus Round – Interpreting Advanced Analytics

If you’ve gone this far, you have made substantial progress toward improving your organization’s data literacy well above par. That being said, if you have a particularly-enthusiastic team, it’s wise to push forward into advanced territory. 

Again, you aren’t training analysts during these sessions, just teaching people to think like analysts. You don’t need to explain how to perform advanced forecasting techniques or prescriptive methods. Instead, simply focus on priming your students to be competent interpreters of the results of those methods.

For example, let’s say they’re going to be using the results of forecasts to determine operations, budgets, etc. They need to understand where forecast methods shine and where they fall short. The objective is for them to be able to make genuinely informed decisions when presented with predictive analytics and other dark magic.

You’re never really done…

Improving data literacy is a long term commitment. You’ll need to continuously evaluate your organization’s data literacy and pivot accordingly. If you’ve made it this far, however, then congratulations!

You may now pass Go. You may collect $200. Heck, buy Boardwalk and Park Place while you’re at it. Go crazy! You and your team have substantially improved your data literacy and can now take on the world!

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.