By: Eric Siegel, Program Co-Chair, Predictive Analytics World for Financial
In anticipation of his upcoming conference presentation at Predictive Analytics World for Financial Las Vegas, May 31-June 4, 2020, we asked Bas Geerdink, CTO at Aizonic, a few questions about their deployment of predictive analytics. Catch a glimpse of his presentation, Fast Data for Finance – How To Apply Streaming Analytics Technology in Financial Decision Making Systems, and see what’s in store at the PAW Financial conference in Las Vegas.
Q: In your work with predictive analytics, what behavior or outcome do your models predict?
A: I work mostly in financial services, where models are use more and more to predict customer behaviour and credit risk. Marketeers want to know where customers are interested in, and this can be predicted by looking at their pageviews and actions in a mobile app. Credit risk predictions focus on the outcome whether (potential) clients are able to afford a loan, and to pay back their monthly fees. I’ve also worked on fraud detection, which is mostly outlier detection.
Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?
A: Predictive analytics is used to group and classify financials transactions. With this information, several use cases are possible such as giving customers insights in their spendings and earnings.
Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.
A: Your AI models are only as good as your data. Before getting top-notch insights and forecasts, make sure that your data storage and ETL pipelines are of high quality. Invest in streaming data ingestions and analysis, since we’re quickly moving from batch to real-time data-driven decisions.
Q: Which technologies do you see taking a rise in the next few years?
A: Streaming data platforms are clearly becoming more mainstream. Cloud is the de-facto infrastructure for all kinds of data storage and analytics loads. Open source will continue to rule, especially big data frameworks (Spark, Kafka, Flink) and model development (Keras, TensorFlow, PyTorch). The data engineering environment will become more mature, with the rise of DevOps and AIOps platforms and tools; this will make it easier to release, update, and monitor software that runs machine learning models in production.
Q: Which skills are becoming important?
A: Data scientists and engineers will remain scarce, so it’s important to train and hire the right people. Roles that guide the process and practice of machine learning will become more common, such as the Analytics Translator and AIOps Engineer. Open source software development skills and model development will be at the heart of many companies; this should not be outsourced or bought.