What Can Businesses Learn from the Sporting Use of Data
Now that evenings are getting longer and there’s a noticeable chill in the air, if you’re like me, you’re spending more time at home in front of the TV cheering for your favorite sports teams. And of course, being a data analytics fan, I’m fascinated with how the sports industry is using data for strategic advantage.
From players wearing connected devices generating data to optimize training regimes in real time, to players using sports equipment with sensors attached to measure swing and range, and soccer clubs using data fueled by AI and machine learning to unearth the next star player, the sports industry is now as much about skill on the field as skill in the discovery and acting upon insights.
But unlike some businesses that have the freedom and time to test and learn using the fail-fast, fail-often and improve philosophy, sports teams don’t have that luxury. You’ve got to get analytics right, especially when you only have one or very few chances to make an intelligent decision. Team managers need to trust the data that they’re dealing with.
Last time, I discussed the issue of data swamps—data lakes that have become dumping grounds for all sorts of data with varying quality. We’ve all heard about the problem of garbage in and garbage out. Like sports teams, you don’t want to make decisions on bad data. And just as an athlete, you want to be able to make intelligent decisions—quickly—for transformative results.
In my past two blog posts, I’ve discussed the advent of data marketplaces—a storefront for trusted high-quality data that curates structured, unstructured, and semi-structure big data for fast analysis. It’s a data lake of pure, fresh water (read: data), rather than a swamp. With data marketplaces, business users have the power to find, access, and prepare enterprise data on their terms, without having to wait for IT.
7 steps to trusted data for analytics
Customers often ask me what the steps are to fuel a data lake with trusted, accessible, and relevant data. All in all, there are a total of 10 steps, but for this post, I’m going to briefly discuss the first seven. You can start small and take each step at a time, but I recommend taking all the steps in sequence.
- Ingest – Data is coming into your business from various places and at varying speeds, so your data management infrastructure should be able to support multi-latency ingestion.
- Stream – The ability to act on insights from streaming data is fast becoming an important competitive advantage. Be sure to take measures to support real-time data that is entering your business.
- Integrate – With data coming through at various speeds and from various places, you need to be able to integrate all types of any volume—at scale.
- Enrich – This is when you infuse your data with expert third-party data, such as business data, so that your analysts can act on rich and updated data. This is when you clean the data as well to ensure data quality.
- Prepare – For extra efficiency, data preparation can be done by data scientists and business users with self-service data prep tools. This accelerates the time it takes to ensure data is fit for analytics and collaboration. It’s also an important foundation for data marketplaces, where users can share the data they’ve prepared.
- Define – At this stage, you define and verify data using data governance policies.
- Catalog – Another foundation to a data marketplace, data cataloging helps data scientists and business users discover, catalog, and curate all enterprise data. This helps users make better informed decisions when they can discover information from around the enterprise.
Here’s a visual that illustrates these steps. You can see there are three other steps after catalog: relate, protect, and deliver. Those are topics for another blog post. If you’re wondering, “CLAIRE” in the graphic is the CLAIRE™ engine, the AI and machine learning technology of our Informatica Intelligent Data Platform.
In the meantime, if you want to learn more about data marketplaces, how they are active data store fronts for big data analytics, and how to build one, download our practical workbook, “Turning a Data Lake into a Data Marketplace.”
Until next time!