Cardinal Analytics

A PRE-ANALYTICS CLEANSING AND MATCHING LAYER FOR INSTRUMENT DATA

cardinal-analytics

Cardinal Analytics is a boutique analytics firm specializing in modelling corporate credit risk. Its cloud-based credit risk information service provides daily rank ordering of bonds for issues  and issuers, allowing traders to analyze the relative risk of their portfolio holdings, generate trading ideas and monitor the market.

Challenge

Three end-of-day data feeds from Mergent, with different naming conventions and identifiers, need to be profiled, cleansed and matched to power Cardinal Analytics’ sophisticated credit risk model along with ratings information. When the firm’s bespoke Python scripts, which required advanced programmers producing thousand lines of code in a 12 week development cycle, failed to reach the goal of ‘near perfect’ match rate, the firm turned to Datactics for assistance to manage both ETL and matching for reference data.

"Every 1% improvement in predictive accuracy creates an additional $1 billion of bond defaults that we can predict."

- Marc Fletcher, CEO, Cardinal Analytics

Solution

Unparalleled Matching Power
FlowDesigner consistently matched 99.5% of bond data (millions of rows of data) achieving the ‘near perfect’ matching power requirements and drastically reducing the need for manual intervention.
Complete Data Transparency
With FlowDesigner, the firm’s data analysts can visualize the match results at each stage of the process and see the effect of changing rules in the underlying data within seconds. If there are any anomalies with the data, FlowDesigner will flag the data to be investigated and corrected in the system prior to being input to the credit risk model.
Ease of Use
Designed for users without programming expertise, FlowDesigner’s intuitive interface with ‘drag-and-drop’ functionality allowed Cardinal Analytics’ data analysts to quickly build ad-hoc processes to meet evolving business needs.

“The Datactics solution is an order of magnitude better than Python code. It is quick andeffective in producing high quality data for the credit risk model, we can see the audit trail all the time, and we can generate high predictive accuracy scores."

- Mark Fletcher, CEO, Cardinal Analytics

 

Cardinal Graph

 

Results

Rapid Replacement of Proprietary System
In just three weeks, FlowDesigner replaced Cardinal Analytics’ black-box Python system, delivering the ‘near perfect’ matching requirements, a fully transparent audit trail and an easily extendible workflow. Cardinal Analytics benefited from a vastly upgraded system that also facilitated the firm’s rapid

Data Quality Assured and Faster Processing Speed
FlowDesigner drastically minimised the manual processing effort needed and provided tools to interrogate and correct anomalies within the system. The quality of data fed into its credit risk model is assured while the pre-analytics process times are also dramatically reduced

Efficiencies and Cost Savings
Valuable technical and management resources that were necessary to maintain the proprietary system and manage the extensive manual review requirements, could be redeployed to core business functions.

 

“With Datactics, we’ve saved valuable resources in terms of maintaining proprietary ETL and matching code in Python and vastly improved our time to market which is crucial in the financial markets. As our company evolves, FlowDesigner allows us to match more companies, immediately add new datasets like non-US bonds and quickly move into other asset classes - essentially the flexibility to extend our product portfolio”

- Mark Fletcher, CEO, Cardinal Analytic

Contact