This is the second and final article in a two-part series that describes the journey of the data science interns Dieudonné Niyitanga and Samuel Mutinda during their placement at a large bank in Rwanda. The blog article describes the overall experience of the two young data scientists with the bank and its data, and the learnings they derived from the process.

Point of departure

By participating in the TA Lite, the bank sought to foster a data-led culture and to create an environment that enables data-led decision-making. The bank realised that it had valuable data assets but lacked adequately skilled staff to ensure these assets were optimally used for business efficiency and profitability. This limited the opportunities to innovate around the data and curtailed the bank’s ability to derive real-time insights and value. Consequently, the bank’s use of data was prolonged, convoluted and limited to departmental reporting of products and services. This is where we, the data science interns, came in. 

An important part of a data scientist’s role is to provide reliable insights, analysis and predictions on risk or future changes in an institution. This empowers an institution to make decisions that can improve business processes and profitability and that can extend products and services to previously untapped or overlooked markets. A data scientist brings valuable skills such as programming, statistical analysis, machine learning and the ability to work with big data to achieve data-driven decision-making.

Our placement, coupled with the support of the data science experts at Ixio Analytics, allowed the bank to refocus its data objectives. This resulted in three successful use cases: a churn analysis, a sentiment analysis via social media and the development of an internal, multi-departmental dashboard. Not only did these projects provide the bank with valuable insights that had been hidden in large quantities of data, but they also taught us a lot about translating data insights into actionable business strategy. We look forward to applying those insights within the bank, as we have subsequently been hired by the bank as their first data scientists on a full-time basis.  

Steering through the bank’s data landscape

We learnt very early in our internships that we had to familiarise ourselves with the state and accessibility of the bank’s data. Understanding the bank’s data management allowed us to plan projects more realistically, as we knew where data access challenges existed and how best to manoeuvre around them. Once we gained the trust of the various departments, we were better able to offer input into more permanent solutions to the challenges we experienced. 

We learnt that our work does not exist in silos and cannot be undertaken in silos, as the projects we implemented were interdepartmental. This demanded that we understand different department objectives and key performance areas to produce more responsive and useful data analytics. Additionally, collaborating with colleagues who understood the business value allowed us to connect the dots between the world of data science and data analytics, and the impact on business value. 

Through its daily operations, the bank collects data from which it can derive valuable insights that support the development of products and services that are more responsive to the market. One of the bank’s requirements of its data was to help it assess which products succeeded and which failed, where services could be improved to increase customer satisfaction, and where costs can be reduced. Having this information at hand allows the bank to make more deliberate decisions that would ultimately reduce the levels of customer churn and increase positive market sentiment. 

However, this is difficult to achieve without developing an operational ecosystem that is receptive and supportive of data scientists. To foster this environment, data scientists must also be malleable and able to translate data science concepts to colleagues from varying professional backgrounds. It also requires that data scientists be willing to stretch their engagement beyond numbers and graphs to meaningful insights that positively impact the bottom-line and increase customer satisfaction. 

It was also very helpful for us to have an internal champion of our work. This supervisor assisted us in our daily tasks and could assist by facilitating engagements with other departments where necessary. Coupled with the technical assistance provided by Ixio Analytics, this internal business support ensured that our work was as relevant as possible to the bank’s business interests and that we had access to the right departments, data and individuals to access resources and ask questions that allowed us to do our work. 

Another important part of the journey for us was learning to translate our data science insights into actionable business insights for the various departments within the bank. Given our limited experience in making business decisions, and the other staff’s limited experience with interpreting large amounts of data, we were often faced with the question, “So what? What does this tell us?” It was crucial for us to continually sense-check what we intended to communicate, as well as engage closely with the various departments to understand what their questions and day-to-day realities were.

Lessons for other data voyagers

If you are interested in what we learned from conducting churn analysis and sentiment analysis for the bank, have a look at the first blog article in the series. If you’re interested in exploring similar projects within your own institution, please reach out to Dumisani Dube at [email protected].

Of course, no two data projects will be the same. Each project will need to be designed in a way that considers the specific objectives, the data assets and skills available, the existing data culture and infrastructure in the institution, and the departments involved. From our experience, it is critical to work with the different departments and to allow enough time to focus on base-line data infrastructure at the start of the project. Having a strong internal champion who has the right level of enthusiasm, technical background and seniority definitely helps there – especially if this is the bank’s first foray into using data science.