Last Updated: October 1, 2024
Fall 2024 Prevedere Platform Updates
We’re excited to share highlights from our latest release. All areas of the Prevedere Platform have new and enhanced capabilities – adding more powerful tools as well as making it easier and more efficient for working with external data and predictive analytics.
Prevedere Global Intelligence Cloud and Discover
Integrating your own internal and other data with our Global Intelligence Cloud is a critical step in building your complete data foundation.
- Data Import Wizard We’ve enhanced the Data Import Wizard to support more data formats, making it easier for you to bring in diverse datasets from various sources. Whether you’re working with proprietary data or external datasets, you can now seamlessly integrate them into the Prevedere Platform, allowing for quicker analysis and decision-making without the hassle of reformatting or manual adjustments.
We’ve made it easier to scale up your analytics and modeling in Discover.
- High Volume Modeling Discover Process Managing multiple forecasts across different variables can be complex, but now you can launch multiple discover jobs at once. This feature is ideal for customers who need to scale their operations, helping you save time by allowing you to create multiple models in bulk, enabling faster analysis across several business units or products simultaneously.
Prevedere Predict
There are a number of new and enhanced features in Predict adding more modeling power and automation.
- Automatic Outlier Detection for Seasonal Profile Creation Outliers can distort your forecasts, especially in seasonal data. This feature automatically detects and adjusts for outliers during the seasonal profile creation process, ensuring that your models remain accurate and reflective of real-world trends, even when unusual events occur.
- Indicator Knitting Economic shocks and structural breaks—like those seen during the COVID-19 pandemic—can disrupt your time-series data. Indicator Knitting identifies these breaks and corrects for anomalies, ensuring your models adapt to these changes. This means you can maintain confidence in your forecasts even when external conditions shift dramatically.
- New Univariate Modeling Methodologies We’ve added 10 new univariate modeling methods to help you find the best fit for your specific needs. Whether you’re forecasting demand, sales, or another key metric, these methods provide a range of options tailored to different types of data and scenarios. Plus, our system automatically recommends the best method based on historical performance, giving you a head start in choosing the optimal model for your use case. Methods include:
- Prophet: Ideal for handling seasonal data with irregular trends.
- ARIMA Airline: Great for datasets with a need for robust trend and seasonal components.
- Holt-Winters (Additive and Multiplicative): Effective for data with consistent seasonal patterns over time.
- Simple and Seasonal Smoothing: Best for simplifying noisy data while preserving trends.
- Complex Exponential Smoothing: Useful for data with both trends and seasonality that changes over time.
- Historic Average: A straightforward approach that uses historical data averages to forecast the future.
- Dynamic Multiplicative Theta: A more complex model that dynamically adjusts based on the dataset.
Prevedere AI Copilot and Assistant
Like all things AI, our AI Copilot and AI Assistant capabilities are rapidly advancing – continuously improving in speed, scope and relevancy – to facilitate faster time to results. Take advantage of your own virtual economist!
- Prevedere AI Assistant Enhancements We’ve made our AI Assistant smarter and faster, based on your feedback. It now delivers more actionable insights with greater speed and clarity, enabling you to find relevant indicators, explain relationships, and derive insights from your models more efficiently. Whether you’re exploring new opportunities or validating a strategy, the AI Assistant is here to help you make data-driven decisions with ease.
- Prevedere AI Copilot Imagine being able to set up complex discover runs just by having a conversation. With Prevedere’s Copilot, that’s exactly what you’ll be able to do. This upcoming feature allows you to interact with the platform via AI chat, guiding you through the discover setup process in a more intuitive and natural way. It’s like having a data science expert on hand to help you navigate the complexities of forecasting, making the experience more user-friendly than ever.
Prevedere Protect
We continue to add new reporting features and visualizations into Protect, as well as optimize the Early Warning System aspects to ensure your external data is constantly monitored and your alerts are timely, informative and actionable.
- Standard Reporting Updates Reporting should empower your decision-making, not slow it down. That’s why we’ve enhanced our standard reporting tools to include new visualizations and easy export options. These updates allow you to communicate insights more effectively to stakeholders, providing clear, visually engaging reports that are ready for business intelligence presentations with just a few clicks.
- Root Cause Analysis Knowing where your forecasts went right or wrong is crucial to improving your predictive accuracy. Our Root Cause Analysis feature dives deep into the assumptions behind your forecasts, distinguishing between statistical anomalies in the model and underlying trends in the data. This insight helps you refine your assumptions, ultimately leading to more accurate and actionable forecasts.
- Model Compare Tab Tracking the evolution of your models is now easier than ever. With the Model Compare Tab, you can quickly analyze multiple versions of a model, compare performance over different time periods, and understand why one model might outperform another. This feature helps you fine-tune your strategies by giving you the ability to compare different approaches side by side.
For More Information
Suggestion: Explore the Prevedere Platform with our Introductory Tour
