Features API - Intro and Getting Started
Access features for your models with our powerful API
What is the Features API?
The Features API changes the way you’re able to use PredictHQ’s demand intelligence. The API aggregates and makes available core statistics from our underlying event datasets, enabling you to use the features directly in your models, without having to do preparatory work to integrate with us. This will significantly reduce the time taken for you to implement our data, and enables us to better serve you.
What are some examples of features?
There are several features we have available for you to try today, these include, by category:
PHQ Attendance: Count, Min, Max, Average, Median, Sum, Standard Deviation
PHQ Rank: Count (Bucketed at a rank level)
Do you want to see the total amount of people attending sports events for 10 miles around a location then you can get that with the features API. Or, do you want to see the number of academic events for a university happening in a time period - you can use the features API for that too. We aim to provide a comprehensive library of different statistics, insights and aggregations on events data.
Where can I try it?
We have in-depth technical documentation and resources such as Jupyter Notebooks available on our tech docs page here.
Also, see our data science guides for guides to using the features API with different types of events (for example features for attended events, for holiday events, for TV events, etc).
Getting started guide
See our getting started guide for examples of how to use the Features API.
The Features API is available within our Python SDK, which you can access here.
As a starting point, we recommend checking out the Features API with PredictHQ’s API explorer available in Control Center. You can use your existing API keys.
See the features API endpoint documentation for documentation on all the fields in the request and response for the features API.
What’s the cost / licensing?
The Features API is currently offered and available on all plans at no extra cost.
How is this different from the Events API?
Our Events API gives event data, Features API provides statistical data. If you are looking for detailed, granular information about specific events, or need that level of detail, the Events API remains our recommendation for this use case. However, if your demand forecasting requirements are more around understanding patterns, trends, and aggregations at a high level across a location, the Features API provides a powerful tool. We encourage you to try both options and seek guidance from our data science solutions engineering team to determine the best API for your needs. For many customers, it will be a combination of both.
What are features?
It’s easy to misunderstand what features are - but it’s really simple in the context of data science, and not to be confused with product features! Features are essentially properties or characteristics from a set of data. This could be a sum, an average, or count of events, across a specific location, or several locations.