Blasting Off Into The Data Universe
As a data driven company, we understand that data insights are essential to optimize and improve our games potential. Dan Fisher, our Senior Data Architect, describes how he created a platform that allows our teams to leverage such vital information.
Soon after joining Outplay I realised a few things;
- There was a large volume of valuable data.
- To capitalise on this, we needed to gather the data in the most accessible format.
- We needed a strategy for getting this information in front of the Outplay team.
Essentially there was a gap in our data and information architecture that needed to be filled to realise its full.
Before designing and creating a solution I decided upon six key principles:
· Data Democratisation: Remove barriers between the users and business information.
· Highest Data Quality: Ensure that the data is always accurate and of the highest quality, with anomalies removed.
· Standardized: Definitions and calculations for key metrics are consistent across the business.
· Aggregated: Support for multi-level aggregations at a company, game and bespoke level.
· Dashboards: Support similar visualisation technologies.
· Technology Agnostic: Portable and interoperable across different technologies.
It was going to be a big project that needed a big name, so the Data Universe was born. It’s first role was to support the creation of a management information system (MIT). The MIT was required to display key performance indicators, allowing a user to apply complex filters, drill downs and drill throughs.
Working together with Pascal Monsewije, one of our data analyst, I set about building phase one of the Data Universe.
Firstly, we modelled how the various events that make up the player journey were related, once we had this in place we could think about the technological requirements. Amazon Redshift was already up and running at Outplay, so presented a lower risk when compared to overhauling the entire technology stack.
We then took the data model and began the creation of a physical implementation within Redshift to hold the myriad of information. During this process we had to do lots of thinking about how the data should be stored, to support aggregations and integration with the visualisation tier. To aid us in this process, Pascal and I arranged various workshops and stakeholder meetings to tease out all the competing use cases and how the data should be displayed and interacted with.
Armed with lots of use cases we set about building the data transformations to extract, transform, and load the data into the Data Universe. I don’t mind admitting that we underestimated this task, mainly because we had to consider today’s data alongside data from legacy systems. Undeterred, we doubled our efforts and persevered!
It was a milestone for us to complete this phase of the project, mainly because it meant that we could have some fun and start to visualise the data. We created lots of concept wireframes before deciding upon a final blueprint. At this point, I finalised the Data Universe, whilst Pascal began building the Dashboards in Tableau and making the data beautiful!
It was a real team effort and good feeling delivering the Data Universe to the business. It’s great to see people from various departments exploring the dashboards and utilising the data to build and improve upon the player experience.
The Data Universe is a great starting point; however, not resting on our laurels, we have already begun forming an ambitious roadmap to take the Data Universe even further in 2018. Plans include building a segmented player CRM and moving to a predictive paradigm, allowing the business to forecast whether a player will spend money or if a player will churn amongst many other things.
As I alluded to at the start, the Data Universe is not tied to a specific technology and to answer the high-value questions and empower the data analytics team, we will be decorating the Data Universe with the latest generation of analytical tools based on Map Reduce and Spark.
The move to a genuine analytics platform allows the Analytics team and everyone at Outplay to contribute to the Data Universe, using smart data science to build insightful data products.
We will be covering the Data Universe more when it moves into Phase Two and other future developments.
