Improper data management represents a significant business risk. And, increasingly, that danger is coming from inside a company’s own walls. Staff members, whether on purpose or by accident, can put their organisation at risk by leaking or making proprietary data available publicly. In the first quarter of 2021 alone, 57 per cent of cybersecurity incidents reported to the UK’s data protection regulator were caused by insiders.

Effective controls on data flows and who can access data are crucial for protecting business information. Refinitiv’s Real-Time Data Access Control System allows companies to manage who is able to access what data, create auditable records of who is able to publish certain types of data, and manage the overall control of it. The system is also able to identify people who have access to certain types of data, but who aren’t actually using it – something that can cut down the risk of data leaks, unintentional or otherwise.

Burak Tutar has a clear idea of what his future looks like: “In five years’ time, our vision is for our company to have fully automated data and execution systems that are maintained by data engineers, where traders are only involved in developing strategies together with quantitative analysts,” says Tutar, who is the founder of AI commodities trading firm Vitus Commodities. 

At the heart of this future vision is the use of artificial intelligence and machine learning. Tutar says the company, which focuses on the energy markets, is using AI to build prediction models capable of forecasting the wind on a medium-term basis. This relies on accurate weather forecasting – something that is notoriously difficult to get right – but also the ability to manage data, utilise cloud computing and deploy machine learning systems on top. 

If successful, it could result in more accurate and cheaper pricing forecasts, automated decision-making for the company and, eventually, automated energy networks. Tutar says that commodity trading houses and the banks of Wall Street are making “somewhat of an effort” to add fundamental data – real-time information about the world, such as satellite images or oil supply levels provided by sensors – into their trading setups, but adds that these are “falling short”. “The main reason for this is simply due to organisational memory and a ‘this is the way we do business’ approach,” he says.

To get to a stage where they can properly utilise the cloud, big data and AI, the organisations need to have the platforms and the requisitely skilled staff in place to do so. Taking in or ingesting data is crucial to this. Refinitiv says its Refinitiv Data Management Solution (RDMS) can provide standardised data that companies can ingest into their own businesses through the cloud, via simple API integrations. It can push all of its data into their data warehousing systems, or help create those systems from scratch.

“We service the developers who want to take in feeds, including real-time feeds,” says Toby Amis, head of developer workflows at Refinitiv. “They want to build internal applications for their own staff.” Kassab adds: “Refinitiv comes with the advantage that it has a huge breadth of content for all aspects of the data that we need to source, all available in one place.” 

Once this data has been brought into a company’s own environment, it can begin to build on top of it. Data can be visualised to show where goods are being moved around the world. Charts can be built to show the world’s busiest ports. Supply levels of commodities such as oil can be analysed – plus artificial intelligence and machine learning can be deployed on this ingested data, potentially unlocking new insights to drive decision-making. 

One Refinitiv project, called Mosaic, uses machine learning to explain extreme price movements. It uses open-source AI technology and adds news sentiment, social media data and information about current events. Hosted on the cloud, the project uses real-time analytics to detect how prices are changing, and a machine learning model then predicts whether the movements are exceptional. To work out whether a price movement is an anomaly, the system has been trained on past trading data, news and previous financial reports. It’s then able to surface these factors to analysts to help them understand what may be causing the price changes. For traders, being able to quickly determine through an automated system why an extreme price movement is happening, allows them to be able to respond and make decisions that will best benefit them and their firms.

Being fluent in the latest technology is critical for staff to stay ahead of competitors

To take advantage of the cloud and AI, companies also need the right staff to be able to utilise them. Data is just one part of AI – and increasingly this means financial firms are employing traders and analysts who have computer and data-science skills, with a particular focus on the Python coding language, which is easier to write than other code and is used in many machine learning systems.

“Many of our customers don't hire people in the front office anymore without Python skills,” Refinitiv’s Amis says. “And that's been the case for some time. Financial services are really adopting Python as a standard.” For companies that don’t upskill their staff, or hire new employees that are fluent in the latest technology, then there’s a real chance they could be overtaken by their competitors. “In five years’ time, if the thought of working in Python or looking at Jupyter Notebook scares people, they're probably not going to have a job,” Amis says.

Kim Benni, the global head of risk and trading technologies at Swiss-based sugar trading specialists Alvean, agrees that Python is a “necessary capability” for many staff these days. However, Benni also points out that there are some artificial intelligence and machine learning environments that are entirely codeless. 

Benni adds that companies being able to innovate with data is crucial – this goes back to the ability to understand business data and have structures in place to utilise this. “The key ability is to understand the underlying business and answer the relevant business questions by
distilling data into insights, producing clean data and usable prototypes algorithms, in a fail early model that promotes creativity and tinkering,” Benni says, adding that companies that look to utilise AI as a tool for understanding trading markets and getting intelligence from data can put themselves at an advantage over their competitors. If they’re not making use of data, then their rivals may be able to forecast what will happen in their own businesses better than they can. “Ultimately,” he says, “No one in trading would even want to be in a situation where competitors can predict your own upcoming sales better than oneself.”

By https://www.wired.co.uk/