Building a Production-Ready Web-Based Application for Real-Time Multi-Scenario Modeling of Investments, Using UxHw® Technology
When making long-term investment plans, organizations and individuals need to make informed choices, even in the presence of uncertainty in future economic conditions. Financial uncertainties, such as future tax rates, future interest rates, and the exact schedule of future investments, all affect the future value of assets. Today, investment advisors and individual savers model the multiple possible future scenarios of their investments (if they do so at all) using Monte Carlo methods; such analyses are too slow to be used often, and certainly too slow to be used interactively.
Signaloid's UxHw®-enhanced computing platforms makes multi-scenario modeling easy to implement and fast to run in real time. Our developer tools, such as the Signaloid CLI, further make it easy to run the code implementing such multi-scenario investment outcome models on the Signaloid Cloud Compute Engine, and the tools also make it easy for organizations to create and deploy web-based applications that provide easy-to-use graphical interfaces for specifying the distributions associated with uncertain inputs (e.g., future interest rates) and displaying the resulting output distributions.
Why It Matters
Across retail banking, small-business and enterprise banking, private banking, and wealth management, customers need confidence in the realism of projections in the face of future macroeconomic uncertainty. Real-time multi-scenario modeling of the future value of investments, which interactively demonstrate the spread and uneven likelihoods of future investment outcomes, give stakeholders a better understanding of how financial uncertainty can affect their investments plans. Signaloid's UxHw technology makes such real-time multi-scenario modeling easy to implement and fast to run. Our developer helper utilities such as the Signaloid CLI make creating production-ready and visually-pleasing interactive applications easy, allowing organizations to deploy production-ready applications in days as opposed to months, and allowing them to increase customer retention, increase revenues, and reduce costs.
The Technical Details
Organizations can deploy their existing financial forecasting model software (compute kernels) on the cloud-based, on-premises, or edge hardware deployment options of Signaloid's compute platform. They can build public or private web-based or mobile applications that send inputs to the software they have deployed on the Signaloid compute engines, and, in running the software, feed in inputs to the software and obtain the output results for display in their applications.
Signaloid's UxHw technology allows the inputs to these deployed compute kernels to contain information about their uncertainty, capturing not just the range or variance but the complete shape of the distribution. For example, an interest rate input into a investment model might be provided as the actual historical distribution of central bank interest rates, assuming the future rates over a long time window will be within the range of historical rates. When run on the Signaloid compute engine, the compute kernels provide the same kinds of distribution output results as compute-intensive Monte Carlo methods, but complete in a fraction of the time. For an investment modeling application we have developed and made open source, the speed running on the the Signaloid Cloud Compute Engine is 580x faster than running on an AWS r7iz EC2 compute instance which has a Intel Xeon Sapphire Rapids processor.
Real applications are however not just compute kernels: They contain user interface elements for providing inputs, interface elements for visualizing outputs, control logic, and more. The Signaloid CLI is a tool we provide for developers to generate complete web-based applications. It takes a compute kernel in C/C++ that can run on the Signaloid Cloud Compute Engine, and provides user interface elements for specifying the input distributions and displaying output distribution results.
One basic example of an investment modeler is a tool for helping individuals estimate their retirement savings based on their monthly savings deposits, interest rates, and tax rates. Because the exact monthly deposits a saver will make for all months until they retire might be impossible to know exactly, it is useful to be able to specify the deposit amounts as a distribution of possible values (e.g., a uniform range between some lower and upper limit, or even as a non-uniform distribution). Similarly, the rates of interest that the savings will be subject to in all future years are often impossible to predict, so these too are useful to be able to describe as a distribution of possible interest rates; the individual or organization interested in the outcome might again choose the distribution to be uniform over a range, or might choose it to match the historical distribution of central bank interest rates over the past several years. Finally, the rates of taxes (if any) that the retirement savings will be subject to are unknowable and would similarly be useful to be able to specify as a distribution of possible tax rates.
As a demonstration of how developers can build and deploy such a multi-scenario retirement investment modeler on the Signaloid Cloud Compute Engine, we've released an example investment retirement account C application as well as a guide for how, using the Signaloid CLI, you can turn that C application into a complete web-based application. The screenshot shows the final polished application, with user interface elements that allow a user to interactively specify distribution inputs, including a "distribution slider" UI element provided by our web-application development libraries. You can try out this completed and polished web-based application on our digital banking industries page.
The Takeaway
When making long-term investment plans, organizations and individuals often need to make informed choices about their investments, even in the presence of uncertainty in future economic conditions. Financial uncertainties such as future tax rates, future interest rates, and the exact schedule of future investments, all affect the future value of assets. Signaloid's cloud-based, on-premises, and hardware module platforms implementing our UxHw technology make implementing and deploying real-time multi-scenario modeling easy. The developer tools we provide, such as the Signaloid CLI, further ease the deployment of production-ready web-based applications built on top of these multi-scenario modeling code kernels running on our compute platforms.
For an investment retirement modeling kernel, running the kernel on the Signaloid Cloud Compute Engine is 580x faster than running on an AWS EC2 r7iz high-performance compute platform. And, you can use the Signaloid CLI to convert an existing modeling code kernel into an interactive multi-scenario web-based application, in minutes.