This site uses different types of cookies, including analytics and functional cookies (its own and from other sites). To change your cookie settings or find out more, click here. If you continue browsing our website, you accept these cookies.
on 08-30-201801:54 PM - edited on 06-19-201911:43 AM by ichand
2018 Excellence Awards: Universal Data Model - Powered by Alteryx
Name: Nikita Atkins
Collaborators: Amazon Web Services
Overview of Use Case:
The Universal Data Model (UDM) is an abstract data model that is designed to capture, clean, integrate,and deliver all insights in simple, flexible, and powerful ways. The UDM allows our wide range clients to accelerate delivery of big data analytics with speed and accuracy thanks to the combined power of a flexible, automated data model and Alteryx Server. Accenture has successfully implemented the UDM in clients across a wide range of industries including Telecommunications, Mining and Retail. Of note Alteryx is at the heart of Accenture’s Sports Analytics Platform being developed in conjunction with a variety of our sporting organisations including Rugby Australia (RA).
Working with coaches, such as Wallabies coach Michael Cheika, and the High Performance Team, Accenture has built a bespoke tool, powered by Alteryx, that will give coaches and players a 360° analysis of their on-field and off-field performance metrics including wellness, biometrics, injuries, training load and detailed play-by-play stats.
Describe the business challenge or problem you needed to solve:
Despite great advanced in technologies and platforms, today’s organisations are still facing the same issues. These include:
Player Benchmarking: coaches need the ability to quickly benchmark and compare players across a range of metrics including game stats, biometric, training load and wellness.
Manual Reports: Sports scientists needed the ability to get standard reports, feedback and insights to players, coaches and staff quicker with less manual intervention. It is estimated that 2 FTE would spend between 1-2 hours developing these reports every day and 3-4 hours for a game day.
Dirty & Siloed Data: Data assets not fully understood, formatting issues & master data not managed in separate systems.
Disparate & Legacy systems: Legacy systems, data warehouse along with large number of spread-marts
Different Business Rules & Requirements: Data requirements & business rules are often different by role, location along with dashboards & advanced analytics
Long Projects & Time to Value: Analytics projects take a long time to deliver business value and often are siloed in nature
No Traceability or auditing:Insights, forecasts and analytics have limited traceability and auditing
Only Quantitative Data:Platforms focused on traditional quantitative data with limited capabilities across documents, and multimedia sources
Data Integration Spaghetti Code: Staging to presentation architectures results in a jumble set of spaghetti code for data integration (ETL, ELT & streaming)
Constant Changing needs:Constant changing needs results in significant amount of rework and additional development.
Describe your working solution:
The UDM and associated sports analytics platform has been developed making use of Alteryx Designer and Alteryx Server to deliver streamlined, repeatable solutions. The UDM is made up of 31 unique macros that handle all aspects of data loading, data cleansing, platform operations, advanced analytics, and auditing. The platform can automatically create feature sets, dimensional models, or data extracts to meet all business needs.
The platform makes use of Amazon S3 and Redshift and deliveries a single view of player, teams and competitors to support all business needs including visualization, reporting, advanced analytics and mobile apps across different data velocities.
Thanks to Alteryx, the UDM and Sports Platform has become highly extensible allowing clients to customize the solution based on their architecture and existing software tools:
The platform makes use of the full range of Alteryx capabilities across data preparation, spatial, in-database, data investigation and predictive.
Parameters drive different behaviors and output for each macro. For example below is a Supertype Load macro that will have different outcomes depending if data input has attributes, spatial, time-series or binary large objects.
For Rugby Australia, the platform has grown at an exponential rate over the last 12 months, capturing every aspect of Rugby played anywhere in the world. Currently the platform has:
4,285 attributes, metrics and statistics for
18,167 unique Athletes playing in
764 unique Teams in
12,867 Matches played worldwide in
58 competitions for the last
16 Seasons capturing
49,774 different in-match Events, along with
35,965 training sessions across
7,341 Locations and Venues, accessible in over
20 dashboards created in Tableau and
1 mobile app
We are now taking the platform to the next level developing sophisticated predictive, machine learning and deep learning models to help improve the performance of Australian athletes and teams on and off field. This has seen the platform expand to include a range of new features including:
Automated predictive model build, validating and testing and pick the best model.
Full auditing and logging capabilities of automated predictive models process.
Creation of new specialized algorithms developed in R (and testing Python) exposed in Alteryx.
Expanded Model Comparison (see below) and Model Score macros making use of new R libraries and new parameters to drive output based on types of model.
Describe the benefits you have achieved:
Below are some of the benefits our clients are seeing making use of our platform.
Specifically, for Rugby Australia, we have seen the following benefits:
Automated creation of dimensional models and feature sets. Changing business rules in the platform take less than 5 minutes and will automatically propagate across all new data and predictive models within 3 hours.
Time taken to create reports by sports scientist has been eliminated. This time is spent creating new dashboards and insights using Tableau on top of the platform.
Coaches, athletes and staff are now connected through a single platform that they can access anywhere, anytime. This means as players go from domestic to international duties that coaches at each level can continue to maintain visibility of their players while ensuring that they continue to remain at optimal health, strength and conditioning.
External website showing the work Accenture has done with Rugby Australia: Link
Ad campaign talking about the work done at Rugby Australia: Link