Construct a customer database from PDF bank statements using Python programming and Microsoft SQL

Abstract

This report proposes a model of extracting customers' transactions information from pdf Bank Account Statement and stores result-set into a customer Microsoft SQL (MsSQL) database for further automated analysis. In nancial sector, it is very important to analysis bank account statement properly to measure the creditwor- thiness for credit approval. To achieve this target, a credit analyst needs to spend a signi cant time for manual analysis which leads to delay credit approval and some- times inaccurate analysis diverts to take wrong approval. So, at present, automated bank account statement analysis is a big demand in the nancial sector. This model will overcome the aforementioned limitations and serve the current market demand. For targeting to achieve this desired goal, the whole process has been divided into 4 basic segments. The rst segment entails converting pdf to text by using a python library (pdftotext), the second one emphasis on correction raw text le (.txt) data by removing unnecessary characters and spaces and do formatting as per need, the third segment consists of parsing formatted text (.txt) and retrieving desired trans- actional information, and nally the fourth segment stores the desired information into a customer database.