PART 1: Data-Driven Decision Making and Analysis at Dyninno Group
Anastasija Grimailova, Data Platform Subdivision Manager, Dyninno Latvia
Welcome to the first part of our series on data engineering and analysis at Dyninno Group. Here, we dive into the critical role of data in modern business environments. We’ll explore how Dyninno Group leverages data-driven decision-making across its diverse sectors, highlighting the importance of data quality in shaping strategic business insights.
Importance of Data in Business Environments
For modern businesses, data is essential across all departments, serving as a significant asset when utilized effectively. This importance spans various levels and teams within an organization, ranging from specialists in local applications like microservices to C-level executives and founders, who analyze daily not only group metrics and key performance indicators, but also operational data.
With the increasing volume of data in every department — such as marketing, finance, and development — the challenges of maintaining data completeness and accuracy are growing, and understanding and managing data becomes crucial. We must focus on data governance, harmonization, and lineage to ensure transparency and clarity behind each metric.
Dyninno Group’s Data-Driven Approach
At Dyninno Group, a large company operating in five business sectors with over 5,100 employees globally, we rely heavily on high-quality digital data analysis for strategic decision-making. Our data-driven approach is integral across all our business divisions. The travel division, Trevolution Group, which operates International Travel Network, ASAP Tickets, Skylux Travel, Dreamport, Oojo, and other travel brands, with over 840,000 unique airline tickets and vacation packages sold in 2023, making it the fourth-largest airline ticket consolidator in the US, for instance, is a significant user of our data platform. While Ecofinance which manages several B2C financial services, CreditPrime, CreditPlus, and Ecofinance, in Romania, Moldova, and Philippines utilizes data ingestion patterns. Currently, we’re expanding our data capabilities to include Multipass, a bank challenger that provides modern financial solutions for businesses with cross-border activity, offering a multi-currency business account with a live FX desk that allows international companies to manage their bank transfers in foreign markets in a simple way.
Our data-centric approach also facilitates machine learning (ML) analysis. Accurate, well-described, and cleansed data is crucial for effective ML applications in our Data Science unit which is currently developing our own in-house artificial intelligence.
Evolution of Data Analysis at Dyninno
The initial phase of data analysis at Dyninno involved forming a dedicated team responsible for data collection and visualization. Initially, our platforms and analytical tools, including a PHP-based BI tool, were developed in-house by our Analytics department.
As our data needs grew, we transitioned to using external platforms for enhanced capacity and efficiency. This shift included partnering with the then start-up Looker platform and migrating data storage and processing to Amazon Web Services. This move marked a significant evolution in our data handling capabilities, aligning with our expanding data requirements.
Data Governance and Machine Learning
It’s crucial to categorize data based on frequency and purpose. At Dyninno Group, some data is processed quickly for daily operational needs, for example, how a marketing campaign is performing currently or how our independent travel agents are operating the incoming requests; while other data is used for deeper, slower analyses or prepared for ML datasets in data science. Developing a strategy for each data type from the outset is recommended, considering the intended use and reasons for data collection.
Data Governance (1) is central to our strategy, encompassing the processes, policies, standards, and practices that ensure effective management of data assets. This framework ensures data quality, availability, usability, and security throughout its lifecycle, positioning data as a strategic asset, 6 main categories have been chosen — Accuracy, Completeness, Consistency, Timeliness, Uniqueness and Validity.
Ensuring Data Quality and Accuracy
As mentioned, the quality of raw data is critical for analysis. Poorly prepared data can lead to significant errors and deviations in results. This underscores the need for qualified specialists to properly prepare data. Taking shortcuts in data management can lead to issues like poor naming conventions, misused properties, or inadequate access control, complicating analyses or causing security breaches, data leaks, and potentially harming the business.
Strategies for Ensuring Data Completeness and Structuring for Analysis
Our current data flow is as depicted below. To ensure completeness and accuracy, we are expanding our Glue options to utilize Data Catalog for both our data lake and data warehouse. At each level, there are various options available for analysis via Glue crawler. Additionally, the Data Quality Definition language in our first MVP is utilized for both technical and business quality checks. We are also currently exploring the integration of Macie for PII data to aid in managing data security and privacy.
For data completeness, we are leveraging technical fields included in libraries within our data collection tool. This allows for increased efficiency in data flow lineage tracking via technical fields.
Data Taxonomy and Business Intelligence Integration
We recognize the importance of structured analytical data in deriving accurate insights. Our use of data taxonomy (2) helps in categorizing and organizing data systematically, facilitating better analysis. A unified view of data across the organization aids in effective data management. This structured approach not only improves analysis but also streamlines access and permission management.
Our adoption of Business Intelligence (BI) (3) solutions democratizes data analysis within the company. By simplifying access to BI tools, we enable all employees — especially in business — to engage in data analysis, irrespective of their technical expertise and daily routine.
To conclude, the foundation of data-driven decision-making at Dyninno Group is robust, but it is just the beginning. Stay tuned for Part 2, where we’ll delve deeper into the transformative steps the company is taking. We’ll explore the cutting-edge of business intelligence systems and forecast the future of data analysis.
(1) Data governance helps make sure a company’s data is good, safe, and easy to use. It sets rules and decides who looks after the data, how to keep it safe, and what it should be used for. The main aim is to keep the data in good shape, secure, and ready for use to understand the business better.
(2) Data taxonomy is a way of organizing and classifying a company’s data. It’s like sorting things into different categories or groups to make them easier to find and use. This system labels and arranges data based on what it’s about and how it’s related to other data. It helps people understand and manage the data better, making it simpler to find the right information when needed.
(3) Business Intelligence (BI) is using data to help companies make better decisions. It involves collecting, studying, and showing data in easy-to-understand ways, like charts or reports. This helps people in the company see trends, find problems, or spot opportunities. The goal is to use this information to make smart choices that improve the business.
