Session Emphases
Full Code
Maximizing Actuarial Efficiency with Python, Git, & DevOps
Presenters: Aaron Smith\Anton Cotelo
This session will demonstrate how Python and Git can be used to simplify and optimize actuarial analysis and operations. Python allows actuaries to efficiently analyze and manipulate large datasets for a variety of purposes and easily interact with a large variety of systems. Git allows for better version control, collaboration, and tracking changes over time of code and other text-based sources. By combining the two, actuaries can gain better controlled processes that are more easily developed and maintained than in the traditional actuarial tech stack. The presentation will also introduce DevOps principles and practices, such as continuous integration, which can further streamline and enhance the actuarial workflow. Attendees will come away with a better understanding of how Python, Git, and DevOps can be used together to improve the efficiency and reliability of actuarial operations.
This is a hands-on demo session, where the ideas will be illustrated with live examples, and participants will be provided with code samples to jumpstart their processes.
So You Think You Can Code?
Presenters: Stephan Marais\Adrian Ericsson
Actuaries around the world are actively looking to make the best of the latest technologies and advances in computing power, data science tools, accessibility to the cloud and machine learning practices. This topic will inform the audience briefly of the different kinds of trending technological and coding projects various actuarial teams are tackling. Knowingly or unknowingly, we start our journey in building and maintaining in-house code, but what are the risks?
We cover an overview of the classic challenges faces by maintaining in-house code from a software development industry perspective. In response to these problems, the software development industry has changed rapidly to attempt to establish “best practices”. Topics such as Agile software development, Test Driven Development, Domain Driven Design, when to use functional vs object orientated programming and continuous delivery will briefly be explained briefly. They are proven to improve long-term code stability, maintainability and speed at which changes can be deployed.
In conclusion, we will look at how these learnings in the software development field could also be employed in the practice of building and maintaining models from an actuarial and statistical perspective.
Actuarial Models & IT: Past, Present, & Future?
Presenters: Pierre MIEHE
New Solvency regimes, new Accounting Standards, new ERM/ORSA frameworks, and shortened delays for closing accounts all put high challenges to actuaries. In parallel, IT and data science / AI are evolving at a huge pace ; but Moore’s law “is dead”*: the improvement of computers performance becomes slower than the growing needs of actuaries. Today, which language to chose to ensure both productivity and performance? Is it easier for an actuary to pick up machine learning / new IT techniques or for data scientists/hackathon specialists to learn insurance? This presentation aims at proposing some answers/trends.
* Jensen Huang, Nvidia Chief Executive, CES Jan 2019”
Low-Code
Unlocking the Power of Data with Power Query: A Low-Code Solution for Actuaries
Presenter: Andrew Chan
In this session, I will demonstrate how Power Query transforms the way actuaries work with data. It enables complex data manipulations without requiring advanced coding skills. As a first-generation low-code data manipulation tool, Power Query is integrated into Excel, providing a versatile and robust solution familiar to actuaries.
With its seamless integration with Excel, Power Query is ideal for actuaries who need to clean, reshape, and analyze large datasets. It simplifies the process of preparing data for complex actuarial models, automates repetitive data processing tasks, and consolidates data from multiple sources, all while maintaining transparency and audibility.
One of Power Query’s key advantages is its ease of learning and use. Actuaries can start transforming data within minutes with no prior programming experience. The intuitive interface and powerful transformation functions allow creating and managing sophisticated data workflows with minimal effort. In addition to Excel, Power Query is versatile across other Microsoft platforms, including Power BI and SQL Server, making it a powerful tool for actuaries working across different environments. This cross-platform functionality scales data manipulation efforts, whether it’s small-scale Excel projects or larger enterprise-level tasks within Power BI or SQL Server.
During this session, I will showcase real-world examples of how Power Query can be applied to everyday actuarial tasks, from data cleaning to advanced calculations. By the end, you’ll understand how to leverage Power Query as a critical component of your actuarial toolkit, maximizing efficiency and productivity in your day-to-day work.
Join me to explore the possibilities with Power Query and see how this no-code tool empowers you to tackle your most complex data challenges easily.
Approachable Automation: A Case Study of Streamlining Actuarial Operations with Dataiku
Presenters: Ande Olsen\Michael Ching
To be successful, today’s actuaries are increasingly required to have advanced data management skills; managing data across numerous interfaces, devising and performing data checks, and performing intricate actuarial workflows are core requirements. These efforts, however, detract from the time available to use their expertise to provide deep analysis and insight.
We will share how we have used the Dataiku Platform to transform actuarial workflows. These improvements have created efficiencies for key activities through automation, reduced risk of error, and improved controls; allowing actuaries to spend more time analyzing results and providing insight to drive business growth.
While we have chosen Dataiku as our Low Code platform of choice, the approaches and capabilities being discussed are transferrable to other platforms. Key elements of our implementation include a unified user interface that bridges multiple applications, real-time documentation of processes, intuitive data interaction with no coding experience required, and a version-controlled and reproducible workflow.
Excel
Building Illustrations in Excel Using Only Dynamic Arrays and Lambda Functions
Presenter: Bryon Robidoux
This session will show how to use the latest Excel 365 features to create a purely functional illustration of a variable annuity policy using only dynamic arrays and lambda functions. This new lambda functionality allows you to create reusable functions in Excel without VBA, which has many advantages. Couple this with Excel 365’s new Excel Lab’s Advanced Formula Environment add-in, and you can now create and download lambda libraries from GitHub Gists. The Gists allow for version-controlled shareable libraries in Excel. The dynamic array functions are designed for ridiculously parallel calculations, but actuarial calculations and illustrations contain highly interdependent calculations. Scan and reduce are the only functions created to manage interdependency, but they have limitations, which I will show you how to overcome with recursion.
As a pricing actuary, I must build illustrations for other actuaries, IT illustration programmers, and admin systems. Classic Excel can hide logic in plain sight with all the complicated formulas and cell references. You can interpret this logic, but it is time-consuming and extremely tedious – especially if it is overly complicated. Given the same set of parameters, pure functions will always return the same result, which makes for much easier testing and validation of Excel calculations. Building a purely functional lambda library makes the logic much more accessible to share with other parties and translate into other languages.
Lambdas make Excel feel more like a traditional programming language, with the advantage of seeing all the intermediate steps. The best of both worlds! The functional programming paradigm is the least popular of the three major paradigms: structured, object-oriented, and functional. The functional paradigm will require you to think about problems differently. But if you are willing to go through this journey with me, the result will immensely help improve your communication of complex calculations with Excel.
Data
Actuarial Data Strategy: Tools, Techniques, and Best Practices in the Era of AI and Next-Generation Tooling
Presenters: William Konop and Jens Christian Wahl
In today’s rapidly evolving landscape, driven by the explosion of AI and next-generation technologies, an effective data strategy is more critical than ever. This presentation explores the essential tools, techniques, and best practices that enable organizations to manage and utilize their data more effectively. Speakers will delve into the use of data compression tools that optimize model run time and cloud cost as well as examine real-world data lake use cases that enable automated and streamlined ETL processes, enhanced analytics, stronger data integrity, and effective data democratization. By integrating these advanced strategies, businesses can stay ahead of the curve, ensuring their data practices are resilient, scalable, and capable of supporting the demands of the AI-driven future.
You will also hear from speakers talk about specific use cases they have implemented with key challenges and issues they have tackled. The live chat functionality will allow you to engage, share ideas, and ask questions.
I Beg To Differ: Actuarial Testware
Presenter: Michael Ching
Day-to-day actuarial work often involves performing arbitrary tabular comparisons. For example,
Systems migration – Checking whether data is identical before and after a systems migration.
Model reconciliation – Ensuring one model’s inputs and outputs match another model’s.
Unit testing – Checking test output during multiple iterations of application development.
Regression testing – Ensuring a code or process change did not impact results.
…and more!
In this session, I’ll introduce I Beg To Differ, a free new open-source application that can help you perform these comparisons. The application features:
Repeatability – Define a comparison once and reuse it again and again!
Multilingualism – Compare SQL Server to Databricks to Excel to CSV to Parquet, and more!
Accessibility – Minimal technical knowledge required!
Automation – An entire Python object model is available for you to automate and extend!
Free up your mental and temporal capacity with this exciting new tool!
Cloud
Cloud Adoption: Empowering Actuaries by Leveraging Cloud Technology for Enhanced Workflows & Reporting
Presenter: Vikas Advani
As insurers search for ways to optimize the value of their actuaries, cloud-based architecture has emerged as a leading solution to deliver performance, automation and network effects that enables the unlocking of actuarial potential. This presentation will cover an introduction to cloud based architecture, actuarial adoption of the cloud, and a live demo of automated actuarial workflow and reporting on the cloud. The demo will also feature how users can access (via a dashboard) live & standardized data from the cloud for ease in reporting. Controls will also be highlighted as a key component of each step of the process.
You will also hear from speakers talk about specific implementations at their organization with key challenges and issues they are tackling today. The live chat functionality will allow you to engage, share ideas and ask questions
Transforming the FP&A Function: Using Automated Model Orchestration and Strategic Forecasting Levers to Improve Efficiency, Empower Users, and Strengthen Controls
Presenter: Jacob Battenberg
This presentation will explore unified modeling, an approach that uses a single automated orchestration engine to centralize model execution across platforms. Unified modeling encompasses automated model runs, processing data for various downstream use cases, and reference data management. Attendees will gain insights into how unified modeling can transform actuarial processes, including its benefits and challenges. Through diagrams and examples in python, we will demonstrate how a single orchestration platform can streamline workflows, address complex business questions, and scale to an entire organization benefiting both actuaries and executives.
Please remember that the views and opinions expressed in these sessions are those of the speakers and do not necessarily reflect the views or positions of their employers, Full Stack Actuarial, the conference sponsors, or any other entity that they may be associated with.