May 31, 2017 | Written by: David Provan
Share this post:
On Sunday, May 28th Roland Garros (French Open) opened its gates to the latest running of the tournament. As one of the four Grand Slam events in Tennis, there is a keen focus on ensuring the event runs smoothly and provides fans with the experience and insights they would want from such a prestigious event.
This year IBM has provided a solution using Watson Analytics to help the content team understand and review key web metrics to help guide decision making throughout the event. Sports events are unique, they are brief moments in time that create an impression, for the organizer of the event they shine an immense focus on their infrastructure, their fan interactions and experience. A key area of these interactions are the digital platforms through which fans engage with the tournament.
Accordingly, the ability to review, make decisions and pivot is vital when hoping to react to fan demands in an incredibly short timeframe. The content team has a primary focus on creating engaging, thoughtful and timely content, the IBM team wanted to provide a simple way to look at and query web metric information that would allow the team to understand the current landscape and make content decisions going forward.
To aid with this the IBM iX team delivered a Bluemix run data ingest system that provided web analytics data into IBM Watson Analytics for review and discovery. The system architecture is detailed below:
We will dive into sections below but at a high level the solution uses an OpenWhisk timer to kick off an API request to a Data API built by the team. This API queries the Web Metrics system and standardizes that data into a DashDB instance. The team then set-up Data Connect to sync data from dashDB into several Watson Analytics Data sets for analysis and exploration.
OpenWhisk has been used as a simple time-based trigger to start the data collection. We have set-up two triggers. The first runs every 15 minutes to gather real time data for each day, the second runs once at 2AM Paris time to ingest a full data-set for the previous day for year on year comparisons.
Data Connect and DashDB
The Data is stored in DashDB and then we created views for each of the datasets to be exported into Watson Analytics. Data Connect was then set-up to sync those views into the Watson Analytics platform every 15 minutes. Detailed below is the easy to use user interface in Data Connect to connect the two data systems.
IBM Watson Analytics
The content team can then log into the Watson Analytics system and quiz and review pre-defined queries of the data to help them understand traffic patterns both in 2017 and in comparison to 2016. Watson Analytics is also able to detect data types and provide visualizations automatically to aid the understanding of the core data, you can see some samples below:
Inside the Watson Analytics system, the content team can query their web metrics data using natural language questions or select a query suggested by the system. From the query, Watson Analytics detects the metrics, date range, and breakdowns desired by the user. Watson Analytics then suggests several visualizations based on the data types of the query components. The user can then explore the data by applying filters as the visualization updates in real time. See example visualizations below:
Bluemix’s platform allows our team to quickly set-up and test these integrations, it’s really the one of the stand out aspects of working with Bluemix. All of our configuration between Bluemix systems was managed with click to configure options and have proved to be very reliable. This is a step towards us creating a metrics system that can provide insights and discoveries to help our properties make informed decisions. The simplicity of creating the platform allows us to build roadmaps based on client desires rather than technical capabilities.
For the organizers of Roland Garros the eyes of the tennis world will look at them for 2 weeks, thanks to the connectivity of the Bluemix tools IBM are able to help them see what how the world is looking at their digital platforms and how they are able to respond to their demands.
(**People involved in the build included @Gray_Cannon)