 Hello everyone, I'm Saif Niazi and I'm here to give you a talk on a data science and financial institute. So, every bank has two types of data. One is transactional where they maintain your money activities and other is non-transactional data where they maintain your personal information like age, methods, status and similar things. But the local banks dump their non-transactional data into the data lake and they don't use it further. So we made a portal follow me which help us to recommend local merchants and customers. So what's the need of this? So if you understand your customer better then you can retain them and retain your customer is a big thing. So for this POC we use a million row data and for storing and ETL purpose we use Hadoop so you don't have to worry about the cost. So this is the technological stack which we use. So for importing the data from Oracle we use a scope and dump the data into the HDFS into the packet format for the compression. So for querying we use Spark. So Spark SQL is much more faster, even 100 times faster than the traditional SQLs if you have that much of a big data. And whatever the result we have is stored in the edge base in a key value pair. Every key has a value so you just transfer that data into the solar for indexing and whatever the insights we have we just show into the portal. I also add a snippet so that developers understand what I'm doing here. So what are the insights we are talking about? So if you're talking about the customer level, like you know your spendings, right? How much money you spend on travel, how much you spending your money on to the food. But if you want to know how much money people similar like you spend. So when I'm talking about people similar like you that means similar in terms of your age, in terms of your salary, in terms of your medical status and similar filters. So you can see that into this portal and analyze what your spending pattern is and how other spend their money. But if you're talking about the merchant level so you can ask the merchant that this particular term have that particular of market you can spend that much of amount there so you get better results. So this is the high level but if you want to tell the local merchants what you should sell there then it's the next step. Like suppose in India majority of the boys who loves to get a coffee after take a coffee they smoke it's only assumption. So not the coffee merchant nor the smoker merchant knows about the habit but only the bank or the financial institute know this because you just swap your card to payment. So because of that you can tell the merchant if you just give a combo of a coffee or a cigarette then you get more benefit. So you can use the data which is dumped in the data leak into some profitable manner. So in last I want to conclude use data because it's a really important asset for all of us and don't store data only for the storage purpose thank you.