“We used to work in silos. Teams used open-source environments with Python, R, and Spark. The CRM team always used Teradata. However, another area used SAS. The first mission was to integrate everything, so we have only one analytical environment,” said Victor Hugo Marques, senior data science manager.
Marques continued, “we were ingesting a lot of raw data. As Bradesco acquires other banks, this brings a lot of complexity when we think about transactional systems. Bradesco has more than 400 transactional systems. Imagine that converging to data. For one subject area, we may have three or four different systems. This translates into data complexity. If we think about databases, tables, and data volumes—it is a lot. For example, mobile, by itself, creates more than 300TBs per day.”