Thanks to Actuarial Society of Malaysia (“ASM”) for inviting ACTBuilder to share our knowledge / research for the benefits and growth of actuarial field in Malaysia. In this “Exploring Data Manipulation & Actuarial Modelling in Python” webinar (18 May 2021), our Principal Consultant (Lim Chee Beng) and key research leader for Python (Lim Jet Kong) shared the outcomes of our initial research in exploring the possibility of using Python in actuarial work. Apart from providing background on Python, they also shared two case studies on movement of policies analysis and actuarial modelling using Python.
For more than a decade ago, actuarial expertise in Malaysia widely used Microsoft Visual FoxPro, specially SQL (Structured Query Language) programming in performing manipulation on policy data and claim data, as well as modelling various calculations. This skill allowed actuarial expertise to perform various analysis on data and design solutions for business problems. However, Microsoft Visual FoxPro was withdrawn from the list of key actuarial tool after Microsoft discontinued to distribute and support this application.
Based on our initial research, we conclude that Python is the potential candidate to replace Microsoft Visual FoxPro as a general purpose programming language for actuarial expertise. Although actuarial expertise are not IT programmers, mastering programming languages strengthen our ability in performing analysis on various data sets – which is inefficient to do it using spreadsheets. In this webinar, Chee Beng and Jet Kong introduced general characteristics of Python and IDE (“integrated development environment”) which promotes a better way in coding. Furthermore, they also emphasized the use of ready-made packages / libraries, such as pandas and NumPy, which can help users to save significant amount of time and efforts – most importantly, they can be downloaded without incurring any costs! For those companies who face constraints in budgets, using Python also helps to reduce spending on software purchase and licensing.
Materials used in the webinar:
- Main slides for Exploring Data Manipulation & Actuarial Modelling In Python
- Sample codes and sample data used to summarize and produce claim development triangles for medical products
- Python Foundation (Download section)
- IDE: PyCharm / Atom / Anaconda
- Programiz online Python compiler
- Google Colab (similar as an online IDE; require login to Google account)
- Python Packages: pandas / Numpy (please note that many packages can be downloaded directly in IDE)
- Python Guides & Tutorial: W3Schools / Real Python