Modelling Data Dependencies

Across the capital markets universe there are a myriad of complex interdependencies always in play. Whether it’s the relationship between an index and its constituents, an option and its underlying or an ETF and its benchmark – there are very few data points that are not intrinsically linked to at least one other (if not many more).

From the perspective of data managers, these interdependencies can be seen as both a curse and a blessing. The curse is that modelling dependencies is a challenge, not least because many data management platforms have been built with little regard for business logic, which means technical barriers can often be a significant hurdle. However, should they be able to overcome that challenge, they will find that modelling data dependencies is in fact a blessing.

Knowing the relationship between different data items can help in a number of ways. For example, if I know that the price of an underlying security has not been validated, then there is no point trying to validate its derivative, which has been calculated from that price. Knowing the nature of certain dependencies can therefore be a key guide in determining workflows. Similarly, if I have just fixed an incorrect closing price for a security, wouldn’t it be great if I could automatically identify all of the data points that depended on that price and will also need updating?

Modelling dependencies can also help craft more accurate data validation rules. Take an index and its constituents; or an ETF and the index that it is designed to track. Modelling those relationships can be vital when it comes to accurately alerting data managers to potential anomalies.

At Xenomorph we have always taken an object-oriented approach to our data model. In doing so, we have been able to flexibly incorporate business rules and define data dependencies to help optimise our clients’ workflows. Having that kind of approach ultimately reaps much greater rewards in the long term.