How to Avoid Abstraction or Generalisation with Remco Broekmans
Executive Summary
When it comes to data modelling, understanding abstraction and generalisation is crucial. This includes analysing information structures, modelling examples, and dealing with communication challenges in business. It's also important to recognise the significance of abstraction in writing and how it relates to healthcare concepts like the "bed". Hospitals must include availability calculations in their models and understand the concept of a physical bed. Consistency is key when abstracting and generalising diagnoses and roles in modelling assets. Business concept hierarchy and terminology must be clarified for effective communication. This includes modelling roles and types in generalisations, analysing language and concepts for organising data and balancing adaptation with industry models. A clear business glossary is essential for communication. Data modelling is essential for businesses, and streamlining the approach is necessary. The benefits of generalisation and abstraction in data modelling cannot be underestimated, and engaging with business analysts and building concepts is crucial during data modelling projects. Active listening and data understanding is critical in workshops, and conceptual workshops play a vital role in model development. Finally, effective communication can be achieved through drawings and presentations.
Webinar Details
Title: How to Avoid Abstraction or Generalisation with Remco Broekmans
Date: 06 June 2023
Presenter: Remco Broekmans
Meetup Group: INs & OUTs of Data Modelling
Write-up Author: Howard Diesel
Contents
Abstraction and Generalisation in Data Modelling
Analysing the Structure of Information and Modelling Examples
Abstract Modelling and Communication Challenges in Business
Understanding the Importance of Abstraction in Modelling
Understanding the Concept of a "Bed" in Healthcare
Importance of Including Availability Calculations in Hospitals
Understanding the Concept of a Physical Bed in a Hospital
Consistency in Abstracting and Generalising Diagnosis and Roles in Modelling Assets
Business Concept Hierarchy and Terminology
Clarifying the Use of Terminology in Modelling and Communication
Modelling Roles and Types in Generalisations
Analysis of Language and Concepts for Organising Data
Industry Models and Balancing Adaptation
Importance of a Clear Business Glossary in Communication
The Importance of Data Modelling in Business
Streamlining Data Modelling Approach in Writing
Benefits of Generalisation and Abstraction in Data Modelling
Importance of Generalisation and Abstraction in Data Modelling
Importance of Engaging with Business Analysts and Building Concepts during Data Modelling Projects
Importance of Active Listening and Data Understanding in Workshops
The Role of Conceptual Workshops in Model Development
Tips for Effective Communication through Drawings and Presentations
Abstraction and Generalisation in Data Modelling
Data modelling involves important concepts of abstraction and generalisation, particularly when creating a data warehouse. Abstraction simplifies and focuses on essential aspects of an entity, while generalisation represents a concept at a higher level of abstraction, often depicted as a tree.
Visualising the data model requires determining which tree parts are essential for understanding and discussing the modelled entity. Finding the balance between including the stem, branches, twigs, and leaves is crucial for accuracy.
Abstraction focuses on the essential aspects of an entity while disregarding non-essential details. Simplifying the model makes it easier to analyse and understand the data. However, careful consideration is required to ensure that important information is not excluded.
Determining the level of detail needed in the data model involves identifying what needs to be represented and understood. This includes understanding when to include just the stem when to include branches or twigs, and potentially even including leaves. The level of detail should strike a balance between providing sufficient information and avoiding unnecessary complexity.
Defining the scope of what is being discussed is essential in data modelling to ensure accuracy in representing the specific entity or concept being analysed. Understanding the scope allows for effective communication and ensures everyone clearly understands the subject matter.
Figure 1 What is abstraction in data modelling?
Analysing the Structure of Information and Modelling Examples
Understanding the bigger picture is crucial. Using a forest as a metaphor for organising information, we can examine the different components, such as trees, branches, leaves, and twigs.
It's important to find balance and determine where to start when discussing a forest or a tree, like an elm tree.
To focus on essential information, we can leave out non-essential details. For example, we can use a fraction level to highlight key elements.
Using personal experience, let's explore how information can be abstracted and modelled. A hospital appointment resulted in a diagnosis of a groin rupture. We'll examine the steps taken to abstract and model the information.
Finally, we'll summarise the information into generalised categories: person, event, place, and thing, based on the specifics of the hospital appointment and diagnosis.
Abstract Modelling and Communication Challenges in Business
Two models are being discussed here. The first model is meant to be a high-level representation that businesses and organisations can use. Unfortunately, it's unclear what the model tries to convey or how it relates to specific events and people.
The second model uses a hospital scenario and includes four elements: person, event, thing, and place. However, it still needs to be clarified how this model differs from the first one and what kind of information it's trying to convey.
To clarify the model, it's recommended to categorise events as hospitalisations or appointments and places as hospitals or hospital rooms. This model also introduces the concept of a person, a doctor, a nurse, or a patient, and a thing, which could be a diagnosis or a bed.
It's essential to have a clear model that effectively communicates the organisational context. The current models need more clarity and could be interpreted differently in different domains such as healthcare, finance, or logistics. A different approach is required to improve communication and understanding.
Figure 2 Abstract Modelling and Communication Challenges in Business
Understanding the Importance of Abstraction in Modelling
When it comes to modelling, Remco suggests trying a different approach. Specifically, he emphasises the importance of including specific details in your model. For example, if you're writing up a model about a hospital visit, it's important to mention things like the appointment, the doctor's diagnosis, and any growing medical issues.
At the same time, Remco acknowledges that it's also important to generalise and categorise terms. This might mean identifying events, persons, places, things, or concepts. For instance, when discussing a patient, it's important to include that category as part of the overall context.
One particular example that Remco brings up is the ambiguity of the term "bed" and its significance within a hospital setting. Drawing on his experiences of being hospitalised for a groin rupture, Remco shares some personal insights about the importance of discussing beds and other related topics in healthcare centres.
Figure 3 Example of Understanding the Importance of Abstraction in Modelling
Understanding the Concept of a "Bed" in Healthcare
In the healthcare industry, there is a debate about whether a bed is simply a physical object or a space where a patient is positioned. To model data voltage in healthcare, events and roles such as doctors, nurses, and patients are categorised to determine similarities and differences. From a modelling standpoint, doctors, nurses, and patients are classified as separate business concepts, but they all fall under the category of a person. The definition of a bed in healthcare is under discussion, with considerations including the physical bed, the surrounding physical space, and staff availability. Some argue that a bed should include not only the physical object but also the area for equipment such as monitors and oxygen, as well as staff availability. Clarifying the concept of a bed in healthcare is crucial to determine its scope and functionality, including its potential placement in hallways.
Figure 4 Understanding the Concept of a "Bed" in Healthcare
Importance of Including Availability Calculations in Hospitals
The capacity of a hospital is not only based on the number of occupied beds but also staff availability. Beds that are placed in the hallway are not counted in the calculation unless they are occupied. They are still considered as available space. For four beds to be available, two people must occupy each room. If only one person is in a room, then the number of available beds decreases to two. If two beds are already occupied, the hospital is at 100% capacity, and immediate action must be taken. Being informed and involved helps understand the process and make necessary adjustments within the hospital.
Understanding the Concept of a Physical Bed in a Hospital
During his hospital stay, Remco mentioned a bet and referred to the bed he was lying on. He was taken to the operating room for surgery and returned to bed. Remco explained that a physical bed is where patients are located in the hospital and referred to it as "bad" for easy identification. These beds are assigned barcodes for maintenance purposes to keep track of their location and servicing needs. Remco believes that using a model to represent the concept of a physical bed in a hospital will help others understand the business process more efficiently.
Consistency in Abstracting and Generalising Diagnosis and Roles in Modelling Assets
Remco acknowledges that it's not consistent to generalise the diagnosis into one table while not doing the same for the roles of doctor, patient, and nurse in a scenario. Remco explains that abstracting the diagnosis at a general level is enough for understanding. At the same time, specific identification is required for roles like doctor, patient, and nurse, as a person can have multiple roles over time. Going one level deeper into the diagnosis reveals the various roles of a person, whereas delving one level deeper into compression shows the different aspects of the person involved. Remco gives an example of a similar issue in a financial institution where modelling assets differed between finance professionals and maintenance workers discussing generators and their components.
Figure 5 the Model: Consistency in Abstracting and Generalizing Diagnosis and Roles in Modelling Assets
Business Concept Hierarchy and Terminology
During the discussion, the focus is on the hierarchy of business concepts, with a particular emphasis on assets and software. There is no mention of equipment, as the discussion centres on the software level. Generators and pumps are categorised as equipment in the hierarchy. The organisation's operations are modelled and communicated using business terminology. Questions arise regarding the classification of various rooms, including hallways and operating theatres, and whether they should be considered rooms or classified under specific categories. The answer to the room classification question is consistent with the approach taken for diagnosis and rooms.
Clarifying the Use of Terminology in Modelling and Communication
When discussing surgical procedures, it's essential to use specific terminology for areas such as the common room, operating room, and theatre. It's important to consider when to simplify or abstract models to avoid overwhelming technicalities. Explaining a generic system to users or adapting it for specific contexts can be challenging, so clear communication is crucial. It's vital to ensure that everyone within an organisation has a shared understanding of terminology, such as a 12-month secured fixed-interest balloon payment loan. Clear communication is also necessary when discussing physical beds and their locations in healthcare settings.
Modelling Roles and Types in Generalisations
During a discussion, Remco suggests that while certain levels may allow for generalisations, adding roles and types requires data for better understanding. Remco briefly mentions modelling doctors but decides not to delve into great detail since all doctors fall under the same category. Howard asks John O'Gorman for their views on building a model, specifically regarding generalisation and ontology. John O'Gorman advises finding a balance between abstraction and instances and suggests modelling certain concepts differently in the q6 framework. John emphasises the distinction between a physical bed and a functional class of a bed in q6, highlighting functional values. John appreciates starting with a general model and acknowledges its value in various situations involving events and people.
Analysis of Language and Concepts for Organising Data
The significance of using clear and precise language in business models is emphasised. The concept of functional classes and location classes is introduced as a means of organising data. Whether particular objects, such as beds, belong to specific classes is discussed. It is highlighted that clients need to be comfortable with the language used and understand how it relates to their data. The importance of organising data to make it easily accessible and searchable is also mentioned. It is suggested that including specific relationships, such as the bed being located in a hospital room, can be helpful. Additionally, the impact of industry data models on building enterprise data models is considered.
Industry Models and Balancing Adaptation
Remco understands that conflicts and challenges may arise when an industry data model clashes with an organisation's perspective. Reference models should inspire organisations rather than dictate strict guidelines. Remco shares a personal experience of their organisation being forced to conform to a specific industry model, which caused difficulties due to cultural and historical differences. Organisations adjust industry reference models to align with their own culture and work processes. Organisations within the same industry may refer to the same reference model but interpret and use it differently. Remco values industry models but stresses the significance of balancing conformity and adaptation.
Importance of a Clear Business Glossary in Communication
Effective communication in organisations requires a clear business glossary. Using generic terms can cause confusion and misunderstandings among employees. Technical and business personnel need to have a shared understanding of the terms. To achieve this, it is essential to have a comprehensive model that defines roles, rules, and types within the business glossary. This model should include descriptions, definitions, and attribution. AI-powered tools like ellie.ai can be helpful in creating and integrating business glossaries. Starting with a well-defined business glossary can streamline the modelling process and improve communication efforts. In summary, a clear business glossary is critical for effective communication and modelling in organisations.
The Importance of Data Modelling in Business
To understand the core of a business, it is important to focus on the patient as the central point, as demonstrated in the snowflake model. It's crucial to find a balance between different frameworks, such as the Goldilocks approach for industries like Coke and Pepsi, and a suitable approach for a medical environment, by using abstraction and generalisation. It's also valuable to explore similarities between different industries, such as mining and medical settings, where equipment is similar to patients regarding movement and location. The presentation discussed the benefits of data modelling as a valuable tool for helping organisations and companies. Recognising the significance of diverse opinions and discussions in finding the most effective solutions for businesses is important. When exploring different modelling approaches, it's essential to settle on a logical data model that is applicable to the author's expertise.
Streamlining Data Modelling Approach in Writing
Remco advocates for a simplified approach to classification, with only a few categories or "buckets" to avoid lengthy discussions. While some people use up to 26 categories to categorise events and individuals, he argues that 5 or 6 buckets are sufficient for an organisation. Remco stresses the importance of precise terminology within a business philosophy. He also emphasises the need to avoid excessive categorisation, which can hinder effective comparison and understanding.
Benefits of Generalisation and Abstraction in Data Modelling
It's important to approach data modelling with an open mind and avoid sticking rigidly to preconceived layouts and models. It can be tough for data professionals to keep an open mind when applying models from one organisation to another in the same industry. Workshops and online sessions are great for encouraging collaboration and using templates. These templates are free to use and require only a simple registration process. Excel-based templates are preferred for their versatility, as they can be easily integrated into other tools like Cash Store. During a Q&A session, one audience member asked about the benefits of generalisation and abstraction in data modelling. The primary benefits identified were reusability, ease of storing data without altering the data model and facilitating understanding and flexibility.
Importance of Generalisation and Abstraction in Data Modelling
To maintain a consistent data model, it's helpful to use generalisation and abstraction. This avoids constantly adjusting the data structure for every new input. Adding a type and role is recommended to ensure it fits into the abstract category when loading data. To retrieve specific data from a general set, data filtering is necessary. While physical data modelling is simple due to minimal changes in the structure, analysing and loading data can be challenging. It's important to consider the trade-offs between generalisation and abstraction to balance data modelling. Collaboration between business analysts, data engineers, and data modellers is crucial for successful data modelling workshops. In these workshops, IT-driven participants provide support and understanding, while business-oriented individuals contribute to the business glossary and definitions. Lack of input or agreement from business people can impede effective data modelling and comprehension of the data definitions.
Importance of Engaging with Business Analysts and Building Concepts during Data Modelling Projects
For an organisation to succeed, it's crucial for its members to be engaged and have a clear understanding of business concepts. The main goal of any project is to ensure that the organisation understands its own operations. To achieve this, the involvement of data engineers and business analysts is essential in comprehending business processes and identifying high-level data definitions. Conducting workshops and engaging with people is necessary to gather information and prevent misunderstandings. Business owners are pivotal in ensuring the accuracy of the business glossary and resolving any disputes that arise.
Importance of Active Listening and Data Understanding in Workshops
During workshops, the writer brings “stroopwafels” to engage participants and takes note of all the terms and information shared by the organisation. They believe it's crucial to capture their words directly. Remco recommends using a big whiteboard to facilitate discussions and encourage participation. He emphasises that developing a comprehensive logical data model and categorising the discussed terms takes multiple workshops. Each workshop lasts for 2-3 hours to allow time for definitions, thinking, and intense discussions. He acknowledges that each organisation and workshop is unique, which makes the process interesting. Remco stresses the importance of understanding the data and relationships in follow-up workshops to ensure accurate modelling. To gauge their understanding, he asks for examples of data relationships between appointments and doctors. Before delving into data modelling, the writer focuses on getting to know the organisation, its business concepts and categorising relevant information.
The Role of Conceptual Workshops in Model Development
The initial workshops facilitate conversation and foster a deeper understanding of the topic. In this context, utilising a data analyst is not to teach core concepts but to guide individuals in the right direction. Starting with an utterly conceptual model can be beneficial but may require adjustments when considering current systems. Conceptual workshops aid in stimulating discussions and refining models. Business professionals need to grasp core definitions, and utilising visual representations can aid in maintaining focus during the workshop. These drawings are used to encourage dialogue and enhance comprehension. During model development, it's crucial to consider the organisation's existing and incoming data.
Tips for Effective Communication through Drawings and Presentations
In addition to written information, drawings are helpful for clear communication. It's crucial to ensure drawings are understandable, regardless of artistic style. Effective communication is essential for conveying information, such as scheduling appointments or comprehending technical details. Presentations that incorporate diagrams and drawings can aid in comprehending complex concepts. Organisations may need to experiment with various approaches because there is no single method for achieving effective communication.
If you would like to join the discussion, please visit our community platform, the Data Professional Expedition.
Additionally, if you would like to be a guest speaker on a future webinar, kindly contact Debbie (social@modelwaresystems.com)
Don’t forget to join our exciting LinkedIn and Meetup data communities not to miss out!