HELPING THE OVERSTRETCHED DATA SCIENTIST.
By Stefan Hendrickx, Founder and Executive Director at Ancoa.
Continuing from our article in the Spring issue “Market surveillance: From cost to value creation”, we move our focus to the analytics use cases of Ancoa, the market surveillance and monitoring platform. We describe how the platform is leveraged by our clients to execute specific data mining tasks, helping the front office gain real insight into both surveillance and risk.
So, how does market surveillance relate to quantitative analysis? As Kara Scannell described in the FT in May 2014 in her article “SEC: With the programme”: surveillance of Wall Street and the financial markets at large requires quant skills and high-tech tools in order to reveal market abuse concealed within huge amounts of transactional data. But these quantitative skills, and the data scientists who are able to do this analysis, are scarce and hardly affordable. Many firms, not only financial ones, want to make use of their precious data crunching skills. The goal is to maximise insight by optimising the workflow of your data scientists.
A number of quants/data scientists at sellside firms and exchanges looking to satisfy the increased need for analysis have worked with Ancoa. From this experience we have found that data scientists and quants often spend more than half of their time preparing data sets for analysis. This unfortunate reality makes it particularly difficult for them to find hidden correlations, which then leads to missed opportunities. Automation in data sampling or “scooping” boosts productivity, and this is where Ancoa can work with firms to help data scientists reach their potential. It is possible to scoop tens of thousands of specific data points out of hundreds of millions, and sometimes billions of transactions, with minimal effort.
Ultimately, the aim is to provide a single analytics environment which runs on top of a single data source. Creating a central data repository for market analytics, which stores both current and historical data on the market has several benefits. At a high level, businesses operating this kind of structure become more streamlined. Different departments learn to speak a common language through shared technology, and there is a reduction in IT overheads and duplication in data storage. Most importantly however, data scientists are able to apply fast and automated data scooping across back and front office. Within this framework, data and the associated analytics used for back office market surveillance and front line traders becomes a ubiquitous feature of an organisation’s structure.
There are implementation challenges in taking a ubiquitous approach to data and analytics. Essentially, it becomes necessary to apply the proper level of governance to ensure that individuals whose behaviour is being monitored do not have access to the alerts being analysed by those doing the monitoring. As a consequence, ensuring that the proper level of governance is put in place is essential in preventing the leakage of information. Nevertheless, the benefits of having a central data repository for analytics easily outweighs the implementation challenges which have to be managed.
A flavour of the types of analytics used by exchanges includes:
- Studying market structure
- Impact of rulebook changes and new connectivity services on behaviour across different groups (buyside, sellside, market makers) and individual market participants
- Building order books from proprietary data
- Effect of policies on liquidity, correlations
- Effects of policies on market structure, network analytics
- Behaviour of individual market participants, in relation to position in order book
- Behaviour of groups or market participants’ (buyside, sellside, liquidity providers) relative positions in the order book for a specific security
A single data repository, at firm level, for analytics has additional benefits for a wide range of firms. Market makers are able to extract statistics on
the performance of individual algorithms. Buyside firms get statistics on different sellside venues, on execution quality and look for substantial trades in the market that are not their own. Sellside firms are easily able to measure performance between traders. Since the application of analytics is immensely diverse, and some of the insight that firms attempt to understand is proprietary, Ancoa has an application programming interface (API) that allows firms to develop their own metrics without disclosing them to third parties.
Ancoa’s key focus is contextual market surveillance. As the article attempts to illustrate, using a real-time system capable of capturing and analysing transactional data in a single repository, facilitates analytics and reports as a natural progression. Ancoa has addressed the corporate data governance issue by offering granular control of user rights and roles, and strong encryption practices. This enables best of breed practices of IT governance levels allowing members of staff to access only the appropriate types of data and applications and avoids inappropriate information diffusion across business functions.
ancoa.com
[divider_to_top]
