Exploring the Higher Levels of Bloom’s Taxonomy

Expanding BI Analyst Horizons with AI

In the landscape of Business Intelligence (BI), the integration of AI and advanced reasoning in the enterprise is essential for staying afloat. To fully integrate with AI, traditional BI analysts, accustomed to working with empirical data and generating reports, must now expand their skills to include deeper analysis, evaluation, and creation. This blog will introduce you to the types of reasoning (empirical, inductive, deductive, and abductive) and how they map to the levels of cognitive recognition within an Enterprise Knowledge Graph (EKG).

Please note that this blog is in the context of my book, Enterprise Intelligence (available at Technics Publications and Amazon). You can read the background on the book’s page. Here is a TL;DR which introduces the terms:

• EKG – Enterprise Knowledge Graph
• ISG – Insight Space Graph
• TCW – Tuple Correlation Web
• DC – Data Catalog

Note that if you purchase the book from the Technics Publications site, use the coupon code TP25 for a 25% discount off most items (as of June 25, 2024).

Over the past couple of decades, whenever someone asked me how to stay ahead of automation, I would rather blithely say, “Stay smarter than the machines.” Of course, that’s much harder to do today, perhaps even close to impossible in a few years or maybe decades (or days?). Even if/when the latter happens, we’re still sentient creatures with feelings and goals. At the very least we can use AI as a kick in the ass to build our sentience and sapience to its fullest.

This blog is inspired by a video I watched yesterday, 6 Levels of Learning Every Student Should Master, by Justin Sung. What caught my eye is the caption on the thumbnail: “A genius thinks at level 6”. That is, Level 6 of Bloom’s Taxonomy–one of those terms that has popped up in my life a few times, but I never dug into. In the theme of “staying smarter than the machines”, I’m anxious to hear anything that helps shed light on what could maximize the intellectual value of people in this time of very good, easily obtainable, and continuously improving AI.

Bloom’s Taxonomy for BI Analysts

Bloom’s Taxonomy, a widely recognized framework in education, provides a valuable roadmap for developing and refining these skills. It categorizes cognitive abilities into six levels:

1. Remembering: Gathering and recalling raw data.
2. Understanding: Interpreting data and explaining trends.
3. Applying: Using knowledge to generate reports and dashboards.
4. Analyzing: Breaking down data to identify patterns and insights.
5. Evaluating: Critically assessing data and insights for accuracy and relevance.
6. Creating: Synthesizing information to develop new strategies and solutions.

Traditionally, and unfortunately, many BI analysts spend most of their time on the first three levels—remembering, understanding, and applying—because these tasks have been the core of their role. However, a good portion of BI analysts actually analyze, presenting findings to managers. Of those, only the more select senior analysts might participate in the evaluating and creation process of strategizing.

To thrive in this AI-driven world, it’s now time for everyone to move into the higher-order skills of analyzing, evaluating, and creating … now … while we are still better than AI at those higher levels. This progression enables deeper insights and more innovative problem-solving, ultimately driving more impactful business decisions.

Interestingly, at 13:36 of the video I mentioned, 6 Levels of Thinking Every Student Must Master, Justin talks about the best way to get to level 5 or 6. He stresses 5 (Evaluating) because he says most of us don’t work at level 6 (Creating). He emphasizes starting at level 5 (Evaluating) rather than beginning with level 1 (Remembering) and progressing sequentially. According to Justin, if we start from level 1 and move up, we tend to forget what we learned at the earlier levels. Our brains have an easier time remembering and working with information at higher cognitive levels.

This resonates with me because my book focuses on integrating big concepts (BI, AI, knowledge graphs, semantic networks, data mesh) and understanding the larger picture. Subsequent books will delve into more detailed aspects. Starting at a higher cognitive level aligns with this approach, encouraging BI analysts to evaluate and synthesize information rather than just recalling facts and applying basic concepts. This method promotes deeper understanding and retention, essential for thriving in an AI-driven world where advanced reasoning and strategic insight are crucial.

But before moving on to Reasoning, all the shameless fear-mongering aside (hahaha), this is a very special time in all of history. There might have been other monstrous transitions for humans, triggered from fire, agriculture, the wheel, smelting, the printing press, etc. But they are still very few across hundreds of thousands of years. No one knows how the next five to ten years is going to play out. So the only sensible play is to release your curiosity from its prison and let it run like a wrongly jailed prisoner released after serving twenty years, gobbling down everything it can find.

A couple generations from now, we won’t be referring to a “gold rush”, but to the “AI summer of the 2020s”.

Reasoning: Inductive, Deductive, and Abductive

To fully engage with the higher levels of Bloom’s Taxonomy, it’s crucial to understand the different types of reasoning that underpin advanced analysis and decision-making.

It’s all very confusing. Much of it is because the reasoning types are used by each other in a kind of circular reference and overlapping way, which makes things confusing for encoders. They are not mutually exclusive concepts. For example, inductive inferences are terms and clues in deductive reasoning and abductive reasoning, respectively. I discuss this more in the book.

Inductive Reasoning

Inductive reasoning involves drawing general conclusions from sets of specific observations. It is the process of identifying patterns and making generalizations based on empirical data (BI data sources, in our case). Most machine learning models (excluding more sophisticated models like LLMs) operate on inductive reasoning. They learn from specific data points to predict future outcomes or classify new data. For example, a machine learning model might analyze thousands of past sales transactions to predict future sales trends. Inductive reasoning helps transform raw data into meaningful trends and insights, making it a fundamental tool for BI analysts.

Deductive Reasoning

Deductive reasoning involves applying general principles to draw specific conclusions. It is about starting with a general rule and deducing specific outcomes from it. For instance, as we’ve all heard, if we know that “all men are mortal” and “Socrates is a man,” we can deduce that “Socrates is mortal.” Deductive reasoning allows BI analysts to form coherent narratives by combining multiple inductive insights, helping them understand the bigger picture and make informed decisions.

Prolog (short for “Programming in Logic”) a logic programming language, is an excellent example of deductive reasoning. In Prolog, you define general rules and relationships, and the system deduces specific answers based on these rules. Whenever I’m confused about inductive, deductive, and abductive, I go to my knowledge of Prolog as the baseline.

It’s unfair of me to toss out “Prolog” and not explain it. That’s way beyond the scope of this blog, but here is a very brief introduction. It operates based on facts, rules, and queries.

• Facts: Basic assertions about the world. For example, machine_learning(llm). declares that LLM (Large Language Model) is a type of machine learning.
• Rules: Logical statements that define relationships between facts. For example, advanced_ai(X) :- machine_learning(X), has_large_scale_data(X). states that any X that is a type of machine learning and has large-scale data is considered advanced AI.
• Queries: Questions posed to the Prolog system to infer new information based on the facts and rules. For example, asking ?- advanced_ai(llm). will prompt Prolog to deduce that LLM is an advanced AI based on the given facts and rules.

Prolog uses these elements to perform deductive reasoning, making it a powerful tool for tasks that require logical inferences, such as knowledge representation and problem-solving in artificial intelligence. Go out and learn the basic of Prolog … it’s a philosophy class for coders.

Abductive Reasoning

To me, abductive reasoning is the formation of hypotheses to explain observations and iteratively refining them. Inductive and deductive reasoning are based on correlations, rules, and conclusions already discovered and vetted. However, the data points derived from inductive and deductive reasoning plug into hypothesis we form and test during the abductive process—that is, constructing provable stories of how something came to be.

This type of reasoning is about generating the most likely explanations from incomplete information. Sherlock Holmes famously used abductive reasoning (even though he often referred to it as deduction). He would observe specific details, generate hypotheses about what might have happened, and then test these hypotheses to find the best explanation. In BI, abductive reasoning is crucial for developing new hypotheses and exploring potential explanations for observed patterns, enabling analysts to create innovative solutions and strategies.

Reasoning in a BI Context

Empirical Data: The Foundation

Empirical data consists of the raw facts and figures collected from various sources within the enterprise. This is the data you are already familiar with—sales figures, customer interactions, and other key metrics. Examples include monthly sales data, website traffic, and customer demographics.

Inductive Reasoning: Insights from Visualizations

In the context of BI, this means generating insights from data visualizations and identifying patterns. For example, you might observe that sales peak during holiday seasons based on historical data. These insights are captured by the Insight Space Graph (ISG) and Tuple Correlation Web (TCW), which record the patterns and relationships observed by analysts. Inductive reasoning helps transform raw data into meaningful trends and insights that inform decision-making.

Deductive Reasoning: Telling the Story

In BI, this means combining multiple inductive insights to form a coherent narrative. For instance, you might conclude that an increase in marketing spend led to higher sales by integrating insights on sales trends and marketing activities. This process is facilitated by the Knowledge Graph (KG), which integrates these insights into a comprehensive story. Deductive reasoning allows you to connect the dots between different data points and understand the bigger picture.

Abductive Reasoning: Creating the Story

In BI, this means developing new hypotheses and exploring potential explanations for observed patterns. For example, you might hypothesize that a new competitor entering the market caused a sudden drop in sales, then test this hypothesis with additional data. Advanced analytics and AI tools enable this process of hypothesis testing and refinement. Abductive reasoning is the creative process of developing new explanations and strategies based on your observations and insights.

Applying These Concepts in EKG

Data collection is the first step, where you gather and organize data from various sources using the Data Catalog (DC). Generating insights comes next, where you analyze data visualizations to identify trends and patterns with the help of the Insight Space Graph (ISG) and Tuple Correlation Web (TCW). Forming narratives involves combining these insights to create a comprehensive understanding of the data, which is integrated into the Knowledge Graph (KG). Finally, hypothesis testing involves developing and testing hypotheses to explain observed patterns, leveraging advanced AI tools and iterative analytics processes.

Expanding the BI Analyst Role

As a BI analyst, your role has traditionally focused on collecting data, generating reports, and providing insights based on empirical data. This typically aligns with the “Apply” stage of Bloom’s Taxonomy, where you use established methods and tools to address specific business questions. However, to stay relevant in an AI-driven world, you need to move beyond this comfort zone and embrace more advanced reasoning techniques. In fact, you must insist on practicing your craft at the higher levels.

Data Scientist Role: A Contrast

Data scientists and BI analysts bring different strengths to an organization, each playing a crucial role in driving business intelligence and strategic decision-making. While data scientists are accustomed to working with complex algorithms, machine learning models, and advanced statistical techniques, BI analysts possess unique skills that complement these technical capabilities.

While data scientists bring technical expertise and advanced analytical capabilities, BI analysts offer crucial business context, strategic insight, and communication skills. Both roles are essential for leveraging data to drive informed decisions and achieve business objectives. By working together, they can maximize the impact of data-driven strategies and ensure that insights are effectively integrated into organizational processes.

Technical Proficiency

Data scientists often excel in exploratory data analysis, building predictive models, and occasionally developing new algorithms to solve complex problems. They are comfortable operating in the “Analyze,” “Evaluate,” and “Create” stages of Bloom’s Taxonomy. This involves abductive reasoning, forming hypotheses, and iteratively refining models based on data.

Business Context and Communication

BI analysts, on the other hand, have a deep understanding of business processes and objectives. They are skilled at interpreting data within the context of business needs and communicating insights to stakeholders in a clear and actionable manner. This ability to bridge the gap between technical findings and business strategy is a significant strength of BI analysts.

Strategic Insight

BI analysts often have a broader perspective on organizational goals and challenges. They are adept at integrating data insights into strategic planning and decision-making processes. This strategic insight allows them to not only analyze and evaluate data but also to contribute meaningfully to high-level business strategies.

Collaboration and Implementation

While data scientists may uncover valuable insights and develop advanced models, BI analysts must play a key role in implementing these findings within the business context. Their expertise in creating reports, dashboards, and visualizations ensures that data-driven insights are accessible and actionable for decision-makers across the organization.

Conclusion

In today’s rapidly evolving business landscape, BI analysts must rise to the challenge of integrating AI and advanced reasoning into their skill sets. Traditional roles focused on data collection and reporting are no longer sufficient. The ability to analyze, evaluate, and create—higher-order skills in Bloom’s Taxonomy—is now essential for staying relevant and impactful.

This blog introduced the types of reasoning—empirical, inductive, deductive, and abductive—and their application within the EKG (which I lay out in Enterprise Intelligence). By embracing these advanced reasoning techniques, BI analysts can move beyond simple data reporting to deeper analysis and strategic decision-making. This shift will not only enhance their capabilities but also ensure that their insights are effectively integrated into organizational strategies.

BI analysts offer unique strengths in understanding business context, strategic insight, and effective communication. These skills are crucial for translating complex data findings into actionable business strategies. While data scientists bring technical expertise, the synergy between both roles can maximize the impact of data-driven initiatives.

The future of BI lies in this integration of advanced AI technologies and the strategic insights provided by skilled analysts. Now is the time to expand your capabilities, embrace higher-order thinking, and make a significant impact within your organization. By doing so, you will not only stay ahead of automation but also contribute to the innovative and strategic growth of your enterprise.