A Simple Guide to Data Literacy

• 70% of employees are expected to work heavily with data by 2025 - Forrester

• 85% of executives believe Data Literacy is going to be the most in-demand skill by 2030 - The Data Literacy Project

The numbers look intimidating and not so far away to their Day 0 impact. However, for any modern day data driven organization, it boils down to 3 key questions -

1. What is Data Literacy by the way?

2. Why now?

3. How to start?

We at Enqurious have been working with some of the world's largest Analytics Consulting companies for the past 3 years helping freshers/laterals get groomed to be deployed at production grade projects faster and better.

The words faster and better may look pretty harmless, even encouraging. However, given the exponential growth of data becoming the core competency to differential businesses, it has become quite evident that working with data is no longer going to be a niche skill.

The world of business witnessed it's first transformational change when office work shifted from paper to software.

It is witnessing it's next transformational change when office work shifts from software to data.

Irrespective of which department you belong to, which background you come from, techie or non-techie, being data driven for decision making is fast taking center stage.

Companies like Amazon has proved time and again that digitizing and measuring each phase of the business has exponential RoI in terms of achieving strategic goals like : Customer satisfaction, Higher operational efficiency and lower financial risks.

So, let's agree to one thing first : It's becoming increasingly difficult to do business just on hunch. Data is taking center stage. So, being data literate becomes the next imperative for organizations.

Great! so, let's address the 1st question :

What is Data Literacy by the way?

Gartner defines data literacy as the ability to read, write and communicate data in context, including an understanding of data sources and constructs, analytical methods and techniques applied, and the ability to describe the use case, application and resulting value.

Further, data literacy is an underlying component of digital dexterity — an employee’s ability and desire to use existing and emerging technologies to drive better business outcomes. 

It simply means if you have data to deal with on your day-to-day life what skills you should have to ensure data is leveraged for operational, financial and business success.

And No! Data Literacy is not about the SQLs, Pythons, Data Lakes, Cloud ETLs, AI/MLs of the world. Tools are the means to empower data driven decision making. However, only a data literate employee would know how best data should be used for sound business decision making.

With the What of Data Literacy clear, let's focus on the 2nd question :

Why now?

Great, so let's move to the final and most important, 3rd question :

How to start?

Let's start with a disclaimer : Data Literacy in its truest sense is much broader than what we are going to present. We are only going to be presenting the how from the perspective of a regular employee who needs to take data driven decisions on a day to day basis using plain Excel sheets.

What we are not going to cover is the governance, cataloging, privacy and security aspects of Data Literacy.

So, let's start with the following illustration :

Let's decode :

1. What's the first thought comes to your mind when you look at some data in Excel?

   What is this data all about? The idea is to first understand the business context, the industry domain the data is all about. The data can be about orders placed by customers in E-commerce industry or claims made by customers in insurance. Without business context, data is simply a bunch of text and numbers. This comes with a follow up question about what business processes exist and how are they measured? Are they digitized yet?

2. What do the bunch of numbers and texts even mean?

   Leave aside different businesses, departments within the same organization interpret same data differently. The same sales column is interpreted differently by Sales team compared to Finance team. Also, when a new request for a report comes, an Analyst faces the next biggest hurdle. How do I even start if I don't what which data is where and what does it mean? This brings out the need for a org-wide, consistent data dictionary. Without a dictionary, we leave data interpretation to tribal knowledge leaving open a wide, risky gap of erroneous outcomes

3. The BI vs Statistics perspective of Data Literacy

   Once the business context and meanings are taken care of, we arrive at a crossroad on whether to analyze data via Business Intelligence approach or Statistics approach? According to us, the answer lies in and not or. Both perspectives lead to a top notch data analysis and an insightful story. Let's look at the two separately

4. The BI Approach

   The world of business intelligence breaks down any dataset into facts and dimensions. Ex. in a sales report, sales amount is a fact which can be measured and aggregated, while dimensions are slicers or perspectives which group/categorize data. Look at the following asks from business :

   1. Prepare a category-wise, state-wise profit report

   2. Prepare a ship mode wise report of deliveries done.

   In the above two requirements, category, state, ship mode are dimensions which group the data and profit, count of delivery is fact. This is important to understand when creating pivots

   However, even before you prepare any report, it is important to understand two ideas :

   1. Granularity of data - What's the lowest level at which your fact is available? Is it country level or country-state level or country-state-district-area level? The 3rd one has lowest granularity of data

   2. Uniqueness of data - How to ensure that data is not duplicated? Are there keys which uniquely define a record?

   Other than this, there are few more thinking points for Analysts :

   1. What new facts or metrics can be derived? Ex. Profit Margin, Return Rate etc.

   2. What new dimensions can be created? Ex. Weekday/Weekend, Month-year etc.

5. The Statistics Approach

   The world of Statistics divides data into 4 types : Nominal, Ordinal, Interval and Ratio. Labels with no order like color is nominal. Ranks or education level is ordinal, Temperature data is interval which sales is ratio. Once this is identified, Analysts are required to profile the data which means to do the following :

   1. Measure Central Tendencies : Mean, Median, Mode

   2. Measure Dispersion : Range, IQR, Variance, Standard Deviation

   3. Measure Shape : Skewness, Kurtosis

   The above steps help understand critical aspects of data like :

   1. If the mean keeps rising, there is a clear upward trend in data

   2. If the mean and median are far apart, there's skewness in data which means there are outliers which need attention

   3. If the spread of data is too sharp, data has high predictability, too flat would mean has high variance, hence low predictability

Finally, this leads to univariate, bi-variate and multi-variate analysis (Not in scope of this article).

Gosh! Isn't that too much for an Analyst to consume to be just called Data Literate?

Well, Data Literacy is not simply a one time course or training. It's a journey which someone genuinely interested in building the skills which would make them highly desirable in modern data driven world.

85% of Chief Data Officers have placed Data Literacy skills to the top spot with a dismal 11% employees using data for decision making.

So, the field is new but one of the most exciting and future defining ones for the knowledge driven economy.

At Enqurious, we help organizations design and deliver a focused and highly effective Data Literacy program. Wish to see how you can build yours?

Connect with us here