Enhancing sports narratives using Natural Language support functionality

How Systech’s deep understanding of data modeling and advanced analytical SQL skills helped the leading entertainment network elevate the quality of LIVE broadcasts for their sports streaming segment. 

Who we worked with: 

One of a mainstream television and film network, reaching over 87 million households worldwide. 

Innovation highlights 

Creation of a sophisticated SQL query analysis model, which can answer a series of questions in near real-time, with 100% accuracy 

Knowledge is power. Continue to discover the entire case study! 


What the customer needed: 

  • A Natural Language Support (NLS) functionality, that can answer questions about their sports segment in near-real time. 
  • Faster and more reliable access to historical and new data, with real-time analysis. 
  • Increased model accuracy that helps to facilitate a more engaging user experience. 
  • The expertise from a network of strong and diverse data scientists to validate the model. 
A demand for a more engaging user experience and stronger advertising 

The sport television arm of a leading network was looking for a way to drive viewer participation, engagement, and usership that could produce stronger advertising and revenue streams.   

The client sought a solution which would assist their sports broadcasters to quickly gain insights throughout the game to help share stories and make comparisons between teams and players to keep viewers engaged. 

To improve the broadcasting experience, the client turned to Systech to create an intelligent data model which could help them synthesize massive quantities of multinational datasets for their sports streaming segment. 

 The objective of this collaboration was to train this model to allow business users to ask a series of sports related questions and get the results in near real-time and help sports broadcasters to share these findings on-air significantly faster and more thoroughly than their competitors. This would in turn elevate the credibility and reputation of the network. 

A need for deep understanding of data modeling and SQL Matching required to achieve lofty analysis goals 

 The client needed to design and execute a sophisticated SQL generator model with its own NLS functionality to offer in-game statistics in real time by tapping into data of athletes, historical game play. 

 Their existing data model lacked an intimate knowledge of linguistics and ability to perform nested SQL queries, which was required to draw parallels between field engines with associative alternatives. Due to the inconsistent and unpredictable nature of the English language, the model was unable to consistently provide results. In turn, the network was wasting the potential of the huge repository of data in their possession for monetary gain. 


 After a few failed attempts from another service provider, the client looked to Systech to help them with all their data and analytic needs and leverage Systech’s expertise and deep understanding of data modeling, SQL Matching, and NLS to optimize a solution for the client. 

Faster and more holistic dissemination of information through a revamped data model 

Systech approached the challenge with a heightened focus on querying linguistics. The team focused on how verbs engaged with the subject, teaching the model how to better adjust to the way human beings maneuver adjectives. At this time in the evolution of data analytics across the industry, it is important that the SQL model had all the parameters, or it simply will not run the query. 

 Data engineers found that while certain sentence constructions resonated with a human audience, the exact same sequence of words would not register. It would also begin to automatically associate different athletes with specific outcomes. For example, if the engineer asked the model which football player had the most running yards, the model may erroneously equate that player with how many yards were run in the game, total. The way the question is phrased affects its ability to process and answer with questions with any sort of reliability. 

 If an effort to combat this underlying threat – amongst others – Systech devised a scoring system to reflect the accuracy and consistency of its data model. Network executives sent an extensive list of questions they wanted answered. After several months of tireless work, Systech was able to perfect a list of over a hundred SQL-compatible, modified questions that could deliver to business users needs with 100% answer accuracy.  

Rapid delivery of sports segment data insights drives tangible revenue in the following ways: 
  • Faster decision making across the business, through deep level question analysis built into the data model. 
  • 100% accuracy across 100+ capable model questions. The increased accuracy of the model empowered the enterprise to drive revenue and advertiser demand through the facilitation of a more engaging user experience. 
  • Faster access to new data, in near real-time, assisting broadcasters in their ability to provide insights on players and historical data in a matter of moments. 

Creation of a more engaging user experience, driving viewer participation, engagement, and usership – producing stronger advertising campaigns and thus a more significant revenue stream. 

  • Enterprise-wide insights to help position better and gain market share in a highly competitive and saturated market, giving them a long-term, sustainable competitive advantage.