This is the repository for the workshop held on 17 Jan 2020 at Beirut Digital District to introduce the Business AI workshops. Attendees learned how to think about AI problems from a business perspective, learned about tools like Power BI, and applied this knowledge in solving a bank marketing business case.
We are introducing you to Power BI as a tool to clean and explore data in a quick and efficient way without any coding knowledge. In order to get you familiar with Power BI, we walk you through a step by step demo using an insurance data set.
Example Power BI visual exploration
Due to recent financial crisis, a bank CEO wants to increase the amount of savings deposit of the existing bank customers base. He asked his Chief Marketing Officer (CMO) to increase the effectiveness of his marketing campaign to convince more customers to put money in a saving time deposited account. The CMO typical marketing campaigns have been based on sending the same marketing message to all the customers in the bank database. The CMO team has 2 marketing resources, a graphic designer and a business analyst in addition to the bank customer support center. The marketing channels that the team used to reach their customers were a combination of phone calls, emails, sms and social media. Recently the CMO has been hearing of advances in AI and data science. However since this is a new strange field to his team, he decided to bring you on board as a consulting team to design new processes that will increase the effectiveness of his marketing campaigns. Ultimately the CMO care about increasing customer life time value by increasing time saving deposits in this difficult period.�
As the team working for the CMO, you have access to data of the 2019 marketing campaigns. The first thing your consulting team does is collect and structure the data into one source of information in the form of one excel table. You set up the data as input fields and output fields. With input being all factors that might possibly affect your output which is “a customer subscribing to a long term deposit”. Assume that each new subscribed long term deposit account returns 200USD/year to the bank. Beside the usual inputs like customer demographics, different campaigns type, call information, you augment the data with socio economic indicators among other variables.
In these tough economic times, you promise the bank CMO that your compensation is tied to a percentage increase in long term deposit plus a small retainer fee conditioned on the CMO implementing your recommendations. You are asked to explore the dataset and answer questions you deem relevant in solving the business problem. You are not required to build an AI predictive model, assume you have it already, and focus on thinking about its integration and implementation from a business perspective.
- Link to the workshop powerpoint slides online
- Power BI demo file
- Power BI demo data file
- Business case data file
- Business case demo power BI file
-
Academic Paper that inspired the business case
Moro, Sérgio, Paulo Cortez, and Paulo Rita. "A data-driven approach to predict the success of bank telemarketing." Decision Support Systems 62 (2014): 22-31.