<aside> 👉🏻 This document discusses the importance of making sense of data in the age of digitalization. It outlines five levels of data quality and emphasizes the need to move from raw data to information and insights through model thinking. The insights gained from building and refining models can help companies create better data strategies and gain a competitive edge.

</aside>

We live in an age of massive data. The advent of micro-processing in the 20th century has spread to all facets of the economy, leading to digitization: the transformation of analog to digital. However, data is an unrefined commodity.

Understand what data you have access to

The first step when working with data is to make a careful inventory of what your company already possesses. Data comes in varying levels of quality:

1. Unstructured data, which is difficult to access. This can be all the data that lives in different files spread out across the organization.
2. Structured and machine-readable data. This can be the contents of spreadsheets, which can be read by another program but that will necessitate specific developments if the data is to be leveraged.
3. Structured data contained in databases. The use of databases makes working with data much easier.
4. Structured, well-documented data accessible through APIs. Data that is accessible through well-documented APIs is very easy to work with and demonstrates significant maturity within the company.
5. Data linked across sets, for context. This is the dream of any Data Scientist. All the data of the company exists in well-organized sets that also contain elements of context in which that data is relevant.

From data to information

While having data can be seen as an asset, it only proves helpful if the company can act on it. But only information should drive our actions - not raw or unrefined data. That is when model thinking comes into play.

<aside> 📐 In model thinking, one uses models - simplified representations of a real-world system or situation - to better understand the underlying dynamics of a system, estimate future behavior, and test hypotheses about how the system might respond to changes in different variables or parameters.

</aside>

By designing and refining models that leverage the data your company possesses or has access to, you can determine what drives systems and behaviors. This is information. This is how data becomes useful.

From information to insights and strategy

Information is what companies rely on to make educated decisions. We use accounting and financial data every day to evaluate the performance of companies.

But building models also provides us with more than just information. We gain insights. Models rarely work out of the gate. They need to be tested, refined, and tested again. Through these iterative stages, it is not unusual to discover:

• missing data, educating the company on what it should aim to capture and collect
• unsuspected variables, hidden drivers that have more impact than you would have ever thought
• uses cases you were not expecting for the data you already possess

These insights enable companies to create better data strategies. Quality in, quality out.