People and technology are the two keys to maximizing the power of data for the organization. This means a growth in organizations hiring data specialists and in analytics and self-service capabilities to extend the influence of data across multiple departments and teams.
Democratizing data across the enterprise can help everyone from the CEO to frontline employees get their jobs done faster, better, more effectively and efficiently. The business intelligence and analytics market is projected to grow by $18.300 billion driven by the need for greater agility, accessibility and deeper insights.
A classic mistake is to think that technology can sufficiently address this challenge, but this only creates drift in initiatives for technological evolution and efficiency. It is not among its functions to promote data literacy and culture per se.
Some advanced analytics solutions could help a new generation of data scientists less technological, but these people come from groups already specialized in statistics and analysis, that is, it does not affect the overall organization.
Some studies predict that 40% of data-related tasks will be automated by 2020, but most of this automation will be applied to the standard loading, cleaning, quality tasks that currently hinder the work of data experts. Even basic descriptive analyzes will not be efficient as they cannot make decisions based on the adequacy of the underlying data.
The educational system will gradually solve the problem of data literacy. While educational institutions understand the need, they continue to fall behind the data literacy needs of the modern workforce. In his 2009 TED talk, mathematician Arthur Benjamin argued that it's time for the math curriculum to shift from analog to digital, and for statistics and probability to replace calculus at the top of the math pyramid.
This necessary change is not in sight in the short or medium term.
It is more urgent to address the talent shortage in data science. It is more worrying that the vast majority of workers are less and less prepared to face this new scenario. Instead of waiting for technology or the educational system to solve the problem, the organization has to analyze, set a strategy, tasks and activate what can be done for data literacy in all areas internally.
For example, a few years ago, marketing departments had almost no data to measure tangible results of their actions. It is now one of the most data-driven and data-strategic departments. To this we must add that they have goals sales and commercial departments.
Keys to data literacy
Data literacy covers a wide spectrum of skills, so it is important to establish a functional base from which to develop specific knowledge for each profile in each case, but with common foundations.
Just as people don't need an advanced degree in English to be literate, people don't need advanced statistical knowledge or programming skills in Python or R to become data literate.
Reading and writing skill levels are often defined by what people can or cannot achieve in their everyday lives; we must do the same with data literacy.
When it comes to basic data literacy, for example, an employee should be able to properly analyze and interpret a standard data chart or graph. They should be comfortable with any of the common charts such as line charts, bar charts, area charts, pie charts, and scatter plots found in most trading applications, dashboards, and news reports today.
Ideally, it would be great if everyone knew how to produce their own charts and do their own analysis, but in my opinion, that's not the minimum standard we're looking for. At a minimum, we need people to be able to consume and interpret data effectively. To do so, they will need skills in the following four areas of data or graphics:
- Knowledge
- Assimilation
- Acting
- Process Review
1.- Knowledge of the data
Each business sector, organization, and department has its own unique set of terms and data. The more employees understand from a business perspective, the better positioned they will be to apply them.
For example, in the case of online sales, you should be familiar with basic metrics like page views, sessions, unique visitors, and bounce rate. In addition to knowing the facts, you need skills in working with numbers and arithmetic.
It is surprising that much of what data scientists focus on is just arithmetic, when the vast majority of analyzes (80%) focus on sums and means. Additionally, a basic understanding of statistical concepts and terms is helpful, such as knowing what correlation is and the difference between quantitative and qualitative data.
2.- Data assimilation
When new data is presented that needs to be interpreted, you have to start the process of understanding unknown data before consuming it. He's not analyzing or making any judgments yet, he's just taking in the information.
For this, the following elements of the tables or graphs must be inspected and clarification requested if any element is ambiguous or a lack of understanding is perceived with certainty:
- Title and labels: Is the table or graph titled and labeled in a descriptive and clear way?
- Time frame: What are the date ranges for the data being presented?
Data Source: Where does the data come from? - Unit(s) of measure: Is it clearly understood what the metrics represent in the charts or graphs?
- Scales: Are the scales of the graph axes clear and effective?
- Calculated Metric(s): For ratios, rates, and other formulas, is there a clear understanding of how they are calculated?
- Dimensions: Are the dimensions or categories used to organize or segment the data clear and meaningful?
- Filters: Is it clear if any specific filters have been applied to the dataset (for example, all customers vs. new customers)?
- Ranking: If securities have been ranked or tranched, is it clear what criteria were used?
- Goals – If goals or objectives have been added to the charts, is it clear what they represent?
3.- Interpretation of the data
Once you are familiar with the data, you can analyze and interpret it. Depending on the type of data and its presentation format, it can be examined in many different ways. In general, you should be used to making the following types of observations on graphs:
- Trend: In which direction is the trend's metric heading (up, down, flat)?
- Patterns: What patterns or repeatable cycles are perceived in the data (for example, seasonality)?
- Gaps: Are there any obvious gaps or omissions in the data set?
- Groupings: Are some values related in certain areas?
- Asymmetry: Are the values noticeably concentrated or skewed more to one side than the other?
- Outliers: Is there a data point that is separate or far away from the rest of the data points?
- Focus: Has anything been emphasized in the graph or table to draw attention to it? Is it obvious why some of the data was highlighted?
- Noise: Is extraneous data included that distorts the main message of the chart?
- Logical: Does the data help answer a specific business question? Do the data support a conclusion or a proposed argument?
4.- Review of the process
In addition to analyzing and interpreting the data, you also have to develop critical and constructive thinking about it. Data is usually accepted as presented. However, it is important to have an overview to assess less obvious factors that may influence the results and their subsequent interpretation.
- Collection method: Could the method or the way the data was collected influence the results?
- Credibility: How credible or trustworthy is the source of the data?
- Bias: Is there a possible bias of the data producer or consumer (including oneself)?
- Truthful: Is data manipulated in a way, intentionally or inadvertently, that misrepresents its true meaning?
- Assumptions: Are there implicit assumptions that could be affecting how the numbers are being interpreted?
- Context: Is there any additional context or background information that is missing and necessary to properly understand the data?
- Comparisons: If supplementary data is included for comparison purposes (for example, period-to-period data), does it provide a fair and relevant comparison? Alternatively, is there an obvious comparison missing?
- Causality: Do you potentially confuse correlation with causation, which represents a direct pattern of cause and effect?
- Significance: If the data is statistically significant, is it also practically significant?
- Outliers: Is it an important outlier or is it unnecessarily biasing the overall results?
- Quality: Can you distinguish between data that is unusable or data that is still useful to leadership or management?
Conclusions
Just as literacy evolves civilizations away from ignorance and poverty, data literacy similarly enriches the organization.
Unfortunately, many users still think that the data is primarily the work of someone else or a more analytical department. In today's data-rich environments, understanding, using and communicating data effectively is everyone's responsibility, not just data experts.
Bridging the data literacy gap in the entity accelerates the ability of employees to take advantage of the increasing amounts of data available.
In addition, it must be taken into account that the repercussions of data illiteracy affect all of society and people, not just organizations. Amid the alarmist cries, fake news, and the proposition of unsubstantiated alternative facts, it is critical to encourage greater data literacy so that people can better distinguish between fact and fiction.
The best protection against misleading data is to immunize yourself against its negative influence through increased data literacy.
