This research work will be published in IJSER (International Journal of Sceince and Engineering Research), 8 August Edition 2012. Part of the research paper wil be posted here:
ELegant Analytics - ELA presents
Personal Finance Intelligence
Medical doctors always say that the best medicine is the preventive one and research shows that humans work very little on this direction. The subject of this research is not Medicine, but not less important, Economy.
Financial crisis happened before, is happening now and will happen in future if we do not create a preventive plan. One of the solutions that aim to be part of a “preventive medicine” for financial crises is produced in the data laboratories of ELA (ELegant Analytics) and its name is PFI.
What is PFI?
Personal Finance Intelligence (PFI) represents the name of the Business Intelligence solution for personal finance and planning of your budget. Inspired by the TV Show “Luksusfellen” (A show broadcasted in Norway about people struggling with their economy and living luxurious live-“Luxurious buddy”.), this Business Intelligence approach may be a solution for all these who fail to maintain well their own economy, for those who want to perform their economy and last but not least the bank itself.
The purpose of this project is to create a Customer Analytical Cube that would process data for each bank costumer using his/her history for its own benefit and then queries back with the most important answers that customers and the bank itself need.
This solution will include also benchmarking against an Imaginary subject that can be Min, Max or Avg of the customer’s values in a set that can be certain region, for a period of time, age group, sex, income ranges etc.
For having more control and planning your own economy, targeting will be an option when users (bank customers) can put their targets for costs or income a month, a quarter or a year ahead and always will be warned when they are about to achieve the costs amount targeted.
Project is also meant to be use for the bank itself in cases when they want to evaluate a customer and his/her behave regarding his/her finance stability, because today Credit scoring system lack for some important data that can make decision more accurate. The customer behave will be same important for the bank, so the bank will know what type of customer is and how he/she handle his/her economy.
Focus and goals
The main focus of this project is the customer, his/her history and his/her behave.
Our first goal is to make possible data collection for customers in the smallest transaction granularity as possible by not impersonating data. This way, bank operates with the whole data diving into details secure and lawfully. Our second goal is by doing Business Intelligence with his historical data to give alerts and advices where he/she is performing bad or giving support where he/she is doing well.
Our third goal is to see where the customer stands, comparing with the region where he/she lives, comparing with his age group, sex and income. This is going to help him improve savings and cut costs by showing how people around him can do with same budget.
Our fourth, but not less important goal is related to the bank itself, where the bank can have clear financial picture for its customer and can decide much better than credit scoring system.
This system lowers the risk and improves the loyalty with customers.
Data center that has the capacity to
• Data transfer once a day or live-data (for the bank side)
• Centralized Customer Intelligence for the Entire Bank
• Live Data transfer and access
• Separate service for each client
• Client vs. Average, Max or Min of a set of clients (Benchmarking)
• Other Intelligence analysis (Geography, age, sex etc…)
Processes on the fly (Administration and Maintenance):
• Optimizing ETL
• Optimizing DB and DWH
• Optimizing Indexing and Data Volume
• Query Performance regarding MDX calculated measures
Project Closure Recommendations
It is very recommended that privacy and security issues regarding credential information about customers to be distributed in a high consideration.
Also users’ impersonation with data source and data reported is highly recommended to be solved in the best possible way, including data source security till Cube role group’s security.
Other recommendation is regarding planning the data volume and performance upon queries requests in a large volume of data in production environment.
Data Architecture of the PFI Solution
* If you want to download the full research just click here.