Tell me about yourself - Senior Data Analyst

I am an experienced expert and a research analyst at CBRT’s Structural Economic Research Department. I am a specialist in real sector analysis, risk, and banking. I am responsible for preparing ad-hoc reports and notes for the upper management about the systemic and credit risks of the financial and non-financial sectors, financial stability, and development issues.

I am also responsible for managing the departmental data warehouse. I was transferred from the Statistics Department five years ago for this and my other capabilities relating to the real and financial sector research and statistics background.

For me, data analysis has two perspectives: researcher and end-user.
I believe a researcher should have an innate drive for data analysis before anything else, and a statistician should have more so. Preliminary data analysis constitutes an essential basis for any type of analysis. With this in mind, I have always valued data wrangling as the first and foremost step for analysis and decision-making. I have spent almost 60 percent of the time throughout my whole working life dealing with all the facets of the data exhaustively. It included all stages from the collection to analyzing and producing reports and further supporting the policymaking.

From a naive user's perspective, on the other hand, data comes last as they usually just want a result that would tell them almost exactly what they need to do. That is to help them make an informed decision. For an analyst, the needs and demands of the end-users are paramount as they determine the schema for data procurement efforts and constrain the data evaluation being part of the establishment of the objective function and targets. Therefore, for an analyst, though a clear-cut definition of the problem and a perfect data source would be the best, rarely do these two come together. Thus, especially data procurement can be defined as a dynamic and recursive process, that demands a continuous back and forth until an optimal model is built; therefore a model-driven one.