The Art of Data

In the 4th century BC, one of the most famous books in the world was written, and despite having been written a long time ago, it remains highly relevant today. It was designed for military strategy but has been adapted to different fields, such as Business Administration. But I wondered, can we adapt this book to data analysis?

The answer was yes, and here I will show you how.

---

1. Know yourself

In The Art of War, Sun Tzu urges us to carefully evaluate our capabilities, resources, and limitations. A detailed understanding of our characteristics is crucial for planning, the first stage in any development or analysis.

Applied to data analysis, we can break it down as follows:

Capabilities

• What knowledge do we have about data analysis? Basic statistics, mathematics, regressions…
• What technologies do we know? Python, Machine Learning, R, Power BI, Tableau…
• What soft skills do we have? Data interpretation, storytelling, effective communication…

Resources

• Servers, IDEs, libraries we know (Python), architecture, ETL tools, etc.

Limitations

• How much data do we have?
• Structure of the available data
• Ease of obtaining new data
• (And we could include here any limitations that arise from analyzing capabilities and resources)

Perfect knowledge of ourselves will help us plan the best strategy to start our analysis.

2. Victory is prepared before the battle

Sun Tzu tells us that preparation for our “battle” is more important than the battle itself. Properly planning our approach ensures a greater chance of success.

• Understand the requirements and the questions the analysis aims to answer
• Data cleaning and preparation
• Invest time in designing the architecture to receive new information, process it, and make it available for correct updating in the corresponding dashboards

This will allow us to design better solutions to answer the right questions that our business seeks to find.

3. Speed and adaptability are key

Sun Tzu tells us that flexibility and speed can overcome a slow and rigid enemy. Being able to adapt to any change and respond quickly can be the key to winning our particular “battle” as analysts.

• Deliver a prototype as soon as possible to receive early feedback
• Use agile development methodologies (yes, they can be adapted to dashboard development teams)
• Deliver periodic report versions to give a sense of constant progress
• Adapt to changes as they arise in the analysis

4. Use intelligence and not just force

Force is fine, but intelligence is better, says Sun Tzu. We must work smart, not harder.

• It’s not about processing more data but focusing on the key information necessary for the analysis
• Sometimes optimization is not about having more resources (hardware), but changing the logic with which we approach the problem (for example, optimizing a SQL query or a Machine Learning model)

5. Avoid unnecessary battles

In the book, a very important part highlights that the best victory in battle is one achieved without fighting.

• The most complex solution is not always the best. If there is a simple solution that tackles the problem as required, that is the best solution.

An example could be choosing a complicated Machine Learning model (for example, complex neural networks) when a simple linear regression could solve the problem.

In summary:

• Know yourself: know your capabilities, resources, and limitations for handling and presenting data
• Victory is prepared before the battle: understand the requirements well and plan your strategy to handle and present data
• Speed and adaptability are key: present constant progress, use agile development methodologies, and adapt to changes in the analysis
• Use intelligence, not just force: it’s about working smart, not harder. Processing more data is not always better. Optimization is key
• Avoid unnecessary battles: the most complex solution is not always the best

Sun Tzu left us an incredible strategy book. Can it be applied to data analysis? Absolutely. The teachings of this strategist could be the key to standing out more as analysts and saving ourselves headaches in future developments and analyses.

What other teachings do you think could be applied?