ASX DataSphere, the Australian Securities Exchange (ASX) data science platform, is now open for third parties for "controlled and governed" collaboration.
Since its launch in 2019, DataSphere has been used within ASX, counting on data that is publicly available. The difference now is that ASX is offering new data sets for the first time and it is making it available to third parties to either use the existing data sets or contribute their own.
Building a marketable data exchange platform
The platform was built over the course of 12 months and counted on technology from Tibco, Cloudera, Talend and Virtustream for the infrastructure, data governance and distribution. Ultimately, ASX built a private cloud for the platform.
This was followed by another six to 12 months when ASX worked on extracting the data from its transactional systems to ensure data governance and safety, until the technology was ready to be used.
"There's a constant process of moving new data onto the platform. We are now getting to a point where we've got a critical mass, if not the vast majority of ASX data on the platform, but it's one data set after the next after the next. So that's a piece of work that's been ongoing for about 12 months as well,” ASX executive general manager for trading services David Raper told Computerworld Australia.
The challenges of building ASX DataSphere
With new technologies coming out all the time and the rapid rate with which existing ones evolve, Raper said that some of the obstacles were around finding the best vendor and suppliers and, after this decision was made, to keep up with the changes and ensure that the right solution was in place.
Although governance and security was always a priority in the project, he said even then it was a challenge to ensure those were “state of the art”. Doing so is not something that comes easily, so the team invested a lot of time and energy into getting it right.
Raper said that the team had a good understanding of the technology itself, so the surprises came from data handling. “I think the surprises really came with the complexity of accessing, understanding, curating and governing data. [Also] there's a lot of work and it always takes a lot longer than you would like or anticipate. It's a voyage of discovery as well: once you actually get that data in place, what can be used, what can't be used, how can it be used, how can it be integrated with other data sets.”
He said that anyone who has done this before would say, “You shouldn’t be surprised by that” but that getting the data organised is a big part of the job. “I'd say that these things always take longer than you hope, and maybe than you plan for. It's important to do them right, because if you have breaches of your data integrity or data security that can really put you back. And because it takes longer than you would anticipate it's very important to have strong sponsorship and organisational support.”
Raper said that it is important to have confidence in the support and backing of the business, but it is just as important to have the patience and belief from your team.
“You can face a lot of challenges, which could force you to cut corners which really doesn't lead to the right outcome.”
Data exchange platforms in Australia
The ASX platform itself is not a first of its kind. Data Republic built its own data exchange platform in 2014 and by 2017 had 50 partners supporting its go-to-market version. Back then, there was interest from the insurance industry as this can be an easy market for fraud attempts.
There were already others in the market with similar propositions but mainly working with publicly available data. Data Republic provides a platform for third parties to securely exchange information. The difference between Data Republic’s platform and ASX DataSphere, according to ASX, is that the ASX platform provides access to extensive and unique ASX data assets.
Alex Linden, an analyst at Gartner, told Computerworld Australia that there are several such platforms in the financial market as data science platforms offer a core of standards, especially around data formats and reuse of data transformation components.
In that market, he said, ASX's platform “can offer optional access to computational resources for smaller investors as added-value services from the ASX,” Linden said. “It offers a set of canned APIs for everybody to use and comply to. This further offers added-value services from ASX, but it can potentially go into the direction of algorithm marketplaces too. It can offer a way to monetise existing proprietary data sets much better. The nice thing about those scenarios is that the raw data doesn’t need to be transferred to a client site any more — it essentially stays inside the data science platform.”
Moutusi Sau, another analyst at Gartner, said that the most value these platforms have created come from market intelligence.
DataSphere’s focus is interest rates rather than the equity or wealth environment, ASX's Raper said. He said that there is already engagement with government agencies such as the Australian Office of Financial Management (AOFM), with the Australian Securities Forum, and with potential consumers or data contributors such as fixed-income dealers, bank treasuries, buy-side investor firms and custodians.
ASX prefers to not quantify its available data in bytes but Raper said the ASX has a data set that, for the first time, will be able to present a whole-market view of the bond market, of the repurchase market and the money market.