The following major activities are covered by our Analytics team:
Data analysis, transformation and mapping:
Data analysis process is critical for every business in order to discover useful information which navigates business decision-making and conclusions. Our company performs sophisticated data analysis activities consistent with our client needs. Our data mining process services includes the following main stages of inspecting, cleansing, transforming and modelling the data. Still, we are not limited to them.
Our experts, and their experience in the field, possess many analytic techniques to turn data into value. They apply different approaches to formulate the right business questions essential for further data sets manipulation and results visualization. Bigger companies, such as corporations, absorb growing volume and variety of data. To help our clients to overcome this challenge, we offer a variety of data transformation solutions which locates the data accurately, in the most appropriate form and at the right time. Further, we perform data mapping process, which includes a wide range of data integration tasks. The data is mediated between data sources and destinations, after that data relationships are investigated and established, which also includes the discovery of hidden data. The final step includes consolidation of multiple data bases into a single one.
Nowadays, credit scoring became one of the most successful and useful applications in banking and finance area used. The most popular type is the application scorecard which produces a score to classify the clients, with great confidence, as ‘good’ and ‘bad’. The scorecard development is a great challenge for each organization in the industry.
Scorecards are well-known and wildly used tools by many of our clients. Their popularity is due to their power in credit risk predictive modelling. EMRC provides great pallet of specific scorecards technology in order to answer our client’s business specifics and needs. We know how to identify and prioritizes the organization objectives for such kind of projects in order to help clients with the competing issues that come up during the development.
In general, scorecard development is a long and sophisticated process, which includes score formulas and score engineering, binning, fitting, objectives and fitting algorithms, characteristic selection, score calibration and scaling, performance inference, bootstrap validation and bagging. Our team has many years of experience in the area during which we have developed many different types of scorecards such as application, behavior, collection scorecards and others. During the years, our company has established expertise in data exploration with information theory, where computation of WOE and IV can become a difficult task. The scorecards can be developed using many techniques and despite the fact that the widely used are discriminant analysis and logistic regression, our capabilities are not limited to them. We are able of applying intelligent system techniques for generating scorecards, such as artificial neural networks and genetic algorithms, which in some cases might be more appropriate.
Predictive Modelling and forecasting:
Predictive Modelling is the essence of a financial institution’s forecasting and consequently for decision-making process in order to make business healthy and sustainable. Our team of professionals provides a broad palette of consistent solutions to improve client’s predictive modelling processes in order to achieve optimal results. Our solutions cover appropriate model technology and model dynamics over time.
Model implementation, calibration and validation:
Recently, it appears that model development is the easiest part of the modelling process. Many of our clients are working in regulated environment and they are aware of the challenges associated with the next steps of model implementation, calibration and validation. In order to provide the most accurate solutions to our clients, we ask the following simple questions: How do you implement the already developed model? Is the model implemented correctly in the company? Are the input parameters and logical structure of the model correctly represented? All those questions are concerning the final target which is to build the right model.
Verification is the essential part where we compare the conceptual model to the computer representation that implements that conception. Model validation is used to determine that a model is an accurate representation of the reality and it is usually done through the calibration of the model. This is an iterative process of comparing the model to actual system behavior and using the discrepancies between the two, and the insights gained, which are used further to improve the model.
Review and Audit:
On-time review and audit are important because performance monitoring ensures model efficacy, enabling our clients to have their best models in production and make consistently better decisions. We are usually involved in the often disconnected array of ungoverned models. Our team observes the model performance, all processes of the model developed, and how they evolve over time in order to determine alerts which show models degrade.
EMRC provides consistent end-to-end governance for all your models and ensures their accuracy status. We manage the company’s workflow and give our clients audit trail management of their models, enabling easier compliance. Our approach consists of monitoring of the model lifecycle throughout the enterprise, which enables managers to easily supervise these models. Additionally, we guide you through the necessary information for regulatory compliance.
Applying new concepts in decisioning:
Each decision in a company is driven by business logic, where the latter is function of the business rules. Recently, in financial industry, taking a decision has become a challenge, mainly because banks and other institutions have to comply with many different rules simultaneously. Some of the rules come from different regulatory institutions, and in certain cases they might even contradict to each other. That is why new concepts in decisioning become so vital. Using our high quality service un the sector, combined with wide experience with different business cases, out clients have already implemented new concepts of decision models to overcome the issues in this difficult and long process.
In general, the decision model orders the business rules into naturally logical groups in order to create a simple structure, which is more understandable, easily navigated and manageable. Our professional decision models give our clients the opportunity to reduce their costs and to increase the quality and effectiveness of the business rules. Additionally, our clients manage to achieve higher level of business understanding of the rules, and simplify business process. Decision Model Policy is so powerful inside the business because it includes governance, control, authorities and designation in one framework.
Bureaux and credit reference agencies link and analysis:
One of the biggest EMRC’s advantages is that during the projects, in which our teams were involved, our professionals have worked with many different data from various data sources, which includes data from Credit Bureau. Our analytical experts have the skills to identify derogatory information, by using data verification tools for potential fraud and further validation.
We have implemented Credit Bureau data for our customers in different projects for acquisition campaigns, decision model development and etc. In these cases, Credit Bureau data is useful because brings account history and customer view across the market, gives more complete view of customer indebtedness, transaction history and performance over mortgages. The data can be further used for verification/ validation of customer characteristics. Additionally, we have used Generic Bureau Scores to determine and implement strategies for new target marketing and product re-launch, for behavior scorecards development, for Risk Capital and monitoring.
Strategic Planning and Analysis:
Competitive and functional strategy.
Six Sigma, DMAIC, DMADV.
PRINCE II, BCPM, SSADM.
analytics case studies
Analytics project: Predictive Modelling and Forecasting Summary
Analytics project: Customer Segmentation and Profiling Summary