Company: American Electric Power
Company background:
American Electric Power is one of the largest electric utilities in the United States, delivering safe, reliable power to nearly 5.5 million regulated customers in 11 states.
AEP owns the nation's largest electricity transmission system, a more than 40,000-mile network that includes more 765-kilovolt extra-high voltage transmission lines than all other U.S. transmission systems combined. AEP also operates more than 223,000 miles of distribution lines. AEP also is one of the nation’s largest electricity producers with approximately 30,000 megawatts of diverse generating capacity, including more than 5,300 megawatts of renewable energy.
AEP's utility units operate as AEP Ohio, AEP Texas, Appalachian Power (in Virginia and West Virginia), AEP Appalachian Power (in Tennessee), Indiana Michigan Power, Kentucky Power, Public Service Company of Oklahoma, and Southwestern Electric Power Company (in Arkansas, Louisiana and east Texas). AEP's headquarters are in Columbus, Ohio.
Contact: Robert Rindos
Title: Data Scientist Associate
Country: United States
Award Category: Innovative Problem Solver
Tell us about the business problem you were trying to solve?
AEP Inc.’s subsidiary, “American Electric Power Energy Partners” (AEP EP), executes large wholesale deals to provide a reliable supply of electricity to their customers. These structured energy deals ensures a stable source of revenue for the supplier and a reliable source of power for the customers. However, counterparty risk is a significant concern in energy transactions, where if the counterparty defaults on its obligations, it will leave the company with losses. To help mitigate the counterparty risk, our team runs analysis for these prospective deals to support credit checks and collateral requirements. The existing process can be a very time consuming effort for each deal valuation, as it requires manual data prep and SAS scripts that include a lot of pain-points. It not only takes our team a few hours to run this analysis, often times the marketers require quick turn-arounds and rely on our analysis to transact deals within a short timeframe.
How did you use SAS to solve that business problem? What products did you use and how did you use them?
We decided to expand and productionize our SAS scripts to become fully integrated within a user-friendly application. We decided to use the R package “Shiny” since our team is familiar with the language and it is fairly easy to get an app running. The R Shiny application, which we call “ProDeals”, is only the start of the process. We designed the app to allow user inputs and parameters, along with a feature that can read and pre-process the structure deal spreadsheets provided by our pricing team. With the use of clever shell and R scripting, we then pass through the pre-processed data and parameters into our SAS programs that will execute on our SAS GRID server. Our pipeline will calculate probability of counterparty default, loss given default, exposure at default based on the inputs provided in the Shiny application. Finally, the completed valuation analyses are populated in real-time in the Shiny application dashboard page.
What were the results or outcomes?
This process would originally take a couple hours to manually set up data, run, and produce analysis for a single valuation. Now we can run multiple valuations at once, all in under 15 minutes. In only a few clicks, we can upload our structured deal spreadsheet inputs into the application, select parameters, run the SAS pipeline, and analyze deal valuations in a comprehensive dashboard. This allows us to provide analysis reports much faster than before, which is essential to keep up with our growing business.
Why is this approach innovative?
The R Shiny integration with SAS has opened up a new door of opportunity for our team. Not only has this greatly improved our ProDeal process, we also plan on developing more apps for non-programmers to use. Apps that are user-friendly, can run SAS programs, and provide analysis on-demand.
What advice would you give to new SAS users?
My best advice is to experiment and think outside of the box. We all will come across tricky problems that can halt our progress, but if you think about the problem in a different way you can sometimes find a unique solution. Sometimes these unique solutions take a lot of iterations and attempts through new technologies or methodologies, but eventually you will learn something new that works. Hopefully you can add that new skill or idea to your toolbox and apply it to your future projects.
