ASX Presses Forward with DataSphere Buildout

While the coronavirus has caused yet another disruption to the Australian Securities Exchange’s planned CHESS replacement project rollout, the operator is still pushing forward with several other major projects.

On Monday, ASX opened up DataSphere to third parties looking to partner with the exchange to solve data challenges. Since launching late last year, the data science platform has expanded the volume of datasets under its umbrella to include data from outside entities. 

Going forward, though, the exchange wants to figure out how to capture and monetize more of the data it creates, says David Raper, executive general manager for ASX’s trading services.

“ASX makes a lot of data available today, so when you see the news that BHP Group has been up three percent, for example, that’s from ASX data. But there’s a lot of other data that is generated within the organization that has never been surfaced,” he tells WatersTechnology.

ASX is working on a range of products via DataSphere that will allow traders, banks, brokers, asset managers, and custodians to understand the volatility of bonds and money market instruments, and the liquidity and concentration risks associated with particular securities. 

The data surrounding the bond, money, and repo markets has traditionally been opaque, Raper says, but the platform is helping to drive some transparency into these realms. “The datasets that we’re surfacing for the first time are by far the most complete view of that market,” he says.

The datasets can be analyzed to provide, for example, a liquidity metric or a concentration risk metric. Then, that metric can be delivered via an API, or integrated into a bank’s risk system, or presented in a visualization application. 

One of the products ASX is currently working on for DataSphere involves corporate actions for bonds. Raper explains that ASX is creating a file that will allow custodians holding bonds on behalf of investors and asset managers to predict and understand the corporate action payments from their holdings, up to 45 days in advance. “This allows them to manage their cashflow more appropriately,” he says. 

Once it’s built for one custodian, it can be easily replicated for other custodians. “It becomes a repeatable product. If it works for Custodian A, it can now be built for B, C, D … They all have a slightly different view because they hold different bonds, but they have the same base requirement of understanding that cashflow. So you essentially work with the first one to refine the product, and then scale it into each individual customer,” he says.

Another example of how ASX is improving DataSphere is around the liquidity of bonds and allowing traders to see how often and how broadly a bond is trading. “It might have traded a lot, but only between two counterparties. Or it might have traded a lot, but with 20 different counterparties, so what’s the concentration? How often are these bonds turning over?” Raper says. “That sort of information that has previously either been completely unavailable or exceptionally difficult [to retrieve] or relied on you being involved in a lot of the transactions yourself.”

DataSphere is built on a customized version of Virtustream, Dell’s enterprise-grade private cloud. The exchange has partnered with open source software integration platform Talend to ingest data from the organization and outside entities. Tibco provides data visualization through its Spotfire platform, and API management through its Cloud Mashery API Gateway platform. Finally, Cloudera serves as the commercial wrapper around the Hadoop and Apache software components in order to run so-called big data analytics. 

Currently, ASX DataSphere has six data products and more than 20 workspace datasets, which are datasets customers can only use within the two workspaces under DataSphere. One is for business and another for data science. Business workspaces allow users to analyze datasets through visualizations and spreadsheets. They can also upload private data. Data science workspaces give data scientists and developers access to analytical tools and allows users to collaborate on product development. 

“Essentially, in a workspace, they can come onto the platform and access R or Python, or whatever else they want to use, and access the data and the compute power of the platform,” Raper says. “And they can also load their own data in there to work alongside ours, if that’s what they choose to do.”

While DataSphere currently has “a few” customers active in the build-out of the workspaces component, ASX is doing most of the legwork, though Raper expects more input from participants in the future. 

“Theoretically, the workspaces are open. So you could take your own workspace and you could do your own analytics completely independent of us, and it’s designed to allow you to do that. You may choose to do that if you have some proprietary data that you wish to analyze alongside ours,” he says. “Practically, however, we’re sort of two weeks into this at the moment, and customers are not doing it independently. They’re coming in and working with us, they’re giving feedback, they’re analyzing a little bit of the data, but then we’re engineering and performing the data science.”

Raper says that it costs “a couple of hundred bucks a month” to get a workspace and the datasets, but reiterates that customers are developing the solution alongside ASX. “So really, the commercialization is less through the workspaces, and more through the subscription to the APIs, the subscription to the visualization, and the subscription to the files. That’s really how it gets commercialized,” he says.