Needing to develop a data-rich software that pulls from multiple sources?
Data Modeling in Finance
FinTech Software that Takes Stock Data and Gives Instant Insights
EquitySet is a FinTech startup headquartered in Chicago, IL that is aimed at providing independent, unbiased stock research to individual investors. Primarily focused on US-based companies, the platform currently covers around 3,000+ major exchange-traded securities.
Beyond equity ratings, the platform provides key financial data and metrics, the ability to track and arrange stocks in watchlists, side-by-side comparison features, and trend tracking among the market and market industries.
An ex-stock and options trader approached LLT with a model and an idea that he had put on the backburner for over 10 years. His vision was to develop a simpler stock rating platform for individual investors that moved away from news-rich and data-heavy platforms like Morningstar or MarketWatch. The financial rating industry was well-established and rooted in a very traditional belief system, so even though the model had been idle the past 10 years, not much had changed.
“The original mantra he brought to us was stock reports that allowed you to get in, get out and get on with your day,” Jake Anderson, Lead Software Developer at LLT Group, said.
He believed that partnering with the right development team could bring the product to market in waves, helping to capitalize on the newfound craze of digitization. The only foreseen issue was that it would take a data-savvy team to help digest, dissect, and deploy all the nuances of what needed to be done. LLT saw an opportunity to help build a very data-rich platform that leveraged the simplicity of a cutting-edge UX/UI to disrupt a complacent market.
Strategy, Experience, Development
Workshop, User Experience, Branding, User Interface, Software
LLT recognized that in order to align the framing around the data modeling, it was necessary to transfer the required knowledge and understanding. In order to achieve this, they hosted several week-long workshops that were on-site with the client to vigorously dive into the algorithms and equations that needed to be converted. They worked side-by-side with the founder to detangle what started out as Excel spreadsheets, written down equations, and outdated examples of the working model.
This discovery period was crucial to not only building the trust needed to form a relationship but also complete most of the heavy lifting for the data modeling. Beyond simply understanding the data points, it was necessary for LLT to find the sources of the data by entering a discovery period where they vetted API data sources to help feed this data-heavy model.
A Development-First Approach for Data
Once they found and vetted all the sources and plugged those into the model, they identified one last, major component – Testing and QA (Quality Assurance). Although LLT knew that the “math” of the model was working properly, they needed to see how it responded to real-world data. This involved the time-intensive, detailed, but significant process of confirming data integrity and consistency from multiple financial data sources.
The development-first mentality was translated across the entire project, first by tackling the data modeling and relationships, then moving on to the features via a crude, functional prototype, but finally completing the process by layering on a platform-based UX/UI.
The final product was a custom-built, Ruby on Rails web application that processes over 40 million financial data points per month based on quantitative models. This also included a background job that functions to not only process the data but also tracks and stores historic ratings across all companies. The platform contains a subscription-based feature that leverages the Stripe API to easily perform account management changes such as free trials, monthly recurring billing, and refunds in cases of subscription termination.
The stock rating features include a fair value stock price calculation along with 9 proprietary metrics that output bullish, bearish, or neutral signals. In addition to the stock modeling, LLT included the buildout of a watchlist feature, a comparison feature, an equity screener, and a sector insight component.
The platform that LLT delivered is finely-tuned machine and is currently entering the market launch phase with the goal of attracting users and doing what it does best – provide independent, unbiased stock reports.