PredictHQ Quarterly Release – October 2022

Published on October 20, 2022

PredictHQ Quarterly Release – October 2022

Welcome to our October 2022 Quarterly Product Release! Each quarter, PredictHQ puts together the latest product and feature announcements to ensure you’re getting the most value out of our platform. You can find previous quarterly releases on our Resources page.

Make better use of our demand intelligence with four new updates that include: 

  • Umbrella events that will help you better understand and prepare for linked events

  • Improvements to our predicted attendance logic for additional categories 

  • Updates to our Location Insights feature

  • A brand new feature engineering guide to help you forecast

Watch the on-demand webinar where PredictHQ Director of Customer Success JJ Mills and Customer Experience Engineer Heath Blandford walk through each of these new releases and share how to use them to get the most out of our data. For now, let’s take a closer look at each of the updates.

Enjoy enhanced accuracy for predicted attendance, powered by umbrella events

Predicted attendance is a crucial part of our data that ensures users are getting the most accurate representation of an event. This quarter's releases include two improvements to predicted attendance: umbrella events and more event category coverage. Let's start with umbrella events. 

An umbrella event is when one event (a child event) belongs to another (a parent event). Our umbrella events feature generates families that automatically link children events with their related parent event. This quarter, we’re excited to announce improvements to umbrella events coverage for conferences, expos, sports, and festivals! This means all events in these categories that run through our pipeline will be analyzed – and if they are part of an umbrella event set, they will automatically be linked together. 

This marks a significant improvement to our predicted attendance feature by ensuring our system avoids double counting of parent and child events. Learn more here, or dive right into the technical documentation here.

Additional improvements to predicted attendance logic – new categories added

When our system collects events that lack attendance data, they’re not useful for forecasting or demand planning. Previously, these data gaps prevented us from predicting accurate ranks and attendance for certain categories of events. 

We’re continuing to fill in the blanks with machine learning models that accurately predict attendance based on similar events–meaning we can now accurately predict the attendance for the events we previously couldn't, including coverage for the following event categories: 

Like our previous installment of predicted attendance improvements, this release enhances our unique ability to better rank events and accurately predict attendance–which no other data source provides.

Updates to Location Insights

Location Insights is a feature in our Web Application (Control Center), where you can easily add your locations of interest to track events impacting them. Since launching the feature, we’ve made several updates to improve the ease of use for saved locations across Control Center, including: 

  • Updated suggested radius feature 

  • Access your saved locations in Control Center search 

  • Daily event impact now supports Location Insights

  • Tooltips for Location Insights stats

Discover the latest updates to Location Insights here, or recap how the feature works here.

Use the Feature Engineering Guide to query and create event-based features

Introducing the Feature Engineering Guide, now available as a notebook in our technical documentation! Data science teams can use this step-by-step guide to:

  • Query various event-based features from the PredictHQ Features API

  • Better understand recommended features per event category

  • Easily create features and include them in your own demand forecasting models, or any other applicable models

The guide includes clear, simple examples to demonstrate which features to use for which event categories. Use the feature engineering notebook to get hands-on experience querying basic features from the features API for: 

  • Attended events (events such as concerts, sports games, conferences, and more) 

  • Non-attended events (such as public holidays, observances and more) 

  • Severe weather events (such as floods, hurricanes, and more) 

Learn more about the new notebook here, or get the guide now, which can be found in the data science section of our tech docs.

Want to dig deeper into these updates? We invite you to check out our October 2022 Quarterly Product Release Webinar. We also encourage you to try the new features out to see how they may improve your workflow or the way you use PredictHQ.