Skip to main content

Data divinity

How we get to the bottom of your data lake

We start by considering if your business goals are aligned with your data and analytics strategy.
In order to achieve this, we use a Data Analytics and AI maturity evaluation model
that consists of the following steps:

The key to building a successful data strategy is to create partnerships between different lines of business. These partnerships are needed to better identify the specific business challenges that can be addressed by data science and analytics.

This entails the creation of a plan regarding the collection, ingestion and storage of data, as well as processing the data in a form that renders it suitable for the creation of new data products.

It goes without saying that in order for value to be generated, companies need to focus on highly valuable data. Hence, data quality and governance play a critical role in the development of your data strategy.

Becoming data-driven embodies not only undertaking technical challenges, but also tackling the organizational aspects related to new functions. It encompasses developing a new data literacy culture within the organization, not just changing the tech stack. Consider the creation of a Data and Analytics COE, as well as roles within the business that can become evangelists and advocates of good data practices (Data Analysts, Data Stewards, Data Scientists, etc.).

How do data scientists collaborate with others in the business to evolve and hand off data science products?

No matter if delivered in an Agile, Scrum, or other way, the methodology needs to start with defining a problem statement; what problem are we trying to solve? Once this is clearly shaped, the analytical approach must be defined. E.g.: Which algorithm is best suited? How can we use the existing data? How can we get a hold of the data we don’t have yet, but will help us solve the problem?

Encompassing all of the above, any methodology should contain solid guidelines for model/product development, and best practices should be followed so that work doesn’t become redundant (feature engineering, model calibration, etc.).

How easily can average business users learn about enterprise data resources? What about Data Scientist – do they have all required data?

How do data analysts and data scientists request and access data? How is data accessed, controlled, managed, and monitored?

How well do the tools used for analytics and data science scale and perform for data exploration, preparation, modeling, scoring, deployment, and collaboration?

When embarking on the journey towards becoming a data driven organization, it’s necessary to start with a thorough analysis of your current data landscape: What has become obsolete? What defers innovation and fosters inertia? Are your current data assets scalable enough? What can be automated? Can non-technical users use data products without a solid programming background?

When your business changes, so must the tools and platforms via which you are trying to reach those business goals.

The current data landscape is evolving at breakneck speed, so make sure you create clear requirements which not only fulfill current expectations but are also scalable and allow you to step into the future.

What tools, including open source, are used within the enterprise for data science objectives? Are they adequate and in line with your business goals or are they obsolete?

How easily can data science work products be placed into production to meet timely business objectives? What obstacles exist in the way of unlocking their full potential?