Fintech is shorthand for “financial technology” and at inception at the turn of the millennium, initially referred to improvements in back-end data storage and processing technology. However, Fintech has since evolved to also include advanced software and algorithms that leverage big data to benefit both financial services providers and their customers.
Matogen AI have successfully concluded a variety of financial services projects, risk modelling in particular, which include developing an online decision science platform. Our expertise in alternative variable creation, extensive data engineering and highly sophisticated modelling techniques have also resulted in groundbreaking methodologies within both the consumer and commercial lending spheres, as well as algorithmic trading.
Disruptive technology
Fintech allows newcomers in the financial services sector to challenge traditional financial operations by offering more flexibility, better and faster services, as well as servicing underserved segments of the customer population. In addition, Fintech includes emerging blockchain technologies such as Bitcoin. Fintech enables conventional financial services providers to improve their product offerings, and hence, profitability. Industry trends include an increase in financial inclusion in the unbanked, developing context, where alternative data (like cellphone use) and gamification is being used to derive credit scores for the unbanked segment, especially in developing economies.
Big data and machine learning
The financial services industry has always been on the forefront of generating “big data”, i.e. data characterised by its large volume, variety and velocity. A lot of traditional financial services data is very structured, automatically generated data, but can also be very unstructured where there was human intervention, for example contactability or employment data. In the highly regulated finance industry, predictive models cannot only be assessed on predictive strength, but also need to be transparent and stable. These well known techniques are then easily transferred to other datasets. For example, it becomes easy to understand logistic regression with weights of evidence transformation.
Risk assessment
Advanced data wrangling, modelling and engineering have especially improved risk assessment, whether in existing credit markets or in market segments where traditional credit profiles are lacking. The use of alternative data has bolstered credit risk profiling for customers with existing credit profiles and has enabled credit scoring for those who have no, or very thin, credit history. This has led to an outburst of innovation in emerging markets in particular, where there is a lack of robust credit registries.
Some of our work in Fintech
Matogen Applied Insights has engaged in a number of cutting edge projects in the Fintech realm such as predicting customer churn, creating alternative features from credit bureau variables, compiling credit scores using alternative data(e.g. Cellphone use data), machine learning for lending policy rules, credit risk incorporating macroeconomic data, as well as engineering an automated crypto currency trading system. The MAI data science team also collaborated on constructing innovative models for clients in the debt counselling sector where customary credit bureau variables are of limited use. In a flagship project, MAI joined forces with a client in the microfinance sector to use psychometric scoring and gamification to determine credit risk. Our services include: