Correlating your data reveals opportunities hiding in plain sight

Go beyond trends and correlate your data with PredictHQ's to identify exactly which kinds of events drive your demand, so you never miss another opportunity.
“With PredictHQ, we have been able to measure the effects of events around the globe in order to discover insights and provide strategic solutions for hotels, restaurants, destinations and retailers.”
Mirko Lalli
CEO & Founder , 
Travel Appeal

The value of establishing correlation

Big data is nowhere near as powerful as smart data. PredictHQ's verified intelligent event data and extensive feature engineering mean you can quickly find your most relevant, high-impact events and clusters of events so you can always be prepared.
View our correlation docs
Python

phq_agg_data = PredictHQ_API(type='aggregate_event_impact')
phq_aggregate_data_df = TransformToDataFrame(phq_agg_data)

model = ForecastingModel()
model.fit(your_demand_data, phq_aggregate_data_df)

future_demand = DataFrame(period=60)
forecast = model.predict(future_demand)
How to use correlation

Correlation with events is complex

Events vary in size, impact, location, frequency and duration, and raw event data is unstandardized and unverified. This makes correlation complicated and time-consuming without PredictHQ.

Let Beam Show you

Beam turns 72 days of work into 52 seconds

  • 01 Collect the data
    Collect and integrate event data into existing dataset.
  • 02 Prepare the data
    Transform your data to match the event data you’re correlating to.
  • 03 Transform the data
    Feature engineer the data. This is an iterative process accelerated by PredictHQ features.
  • 04 Create models
    Create and finetune models for verifying correlation and predicting demand.
  • 05 Take action
    Inform business decisions based on the patterns and demand spikes you detect.

    Identifying a formula for correlation

    Finding the statistical relationship between historical events and transactional demand helps businesses understand event impact. This enables your team to make reliable predictions about how future events will impact business operations and transform upcoming demand fluctuations into revenue opportunities.

    There are many statistical approaches that data science teams take to establish correlation between historical events and transactional demand. Time series modelling is a popular choice used to decompose transactional demand in order to study the relationship between the impact of events and demand fluctuations.

    But transactional demand patterns can be complex and relevant event data can be difficult to source and process, so advanced time series modelling can take months to complete. We created PredictHQ Beam, our correlation engine, to turn months of work into minutes. Take the labor out of establishing correlation so your teams can spend more time perfecting forecasting models .

    Documentation

    A step-by-step guide to establishing correlation and improving your forecasts

    Get practical advice by downloading and reading through PredictHQ’s technical guide.

    Our data pipeline

    Correlation is the foundation for accurate prediction

    Transforming the data is typically the most time consuming and error-prone part of the process. Extracting features from raw data to better represent the underlying problem to predictive models is critical for model accuracy. Better features, means better results.

    Prediction

    Get Started

    Contact our data science experts to find out the best solutions for your business. We'll get back to you within 1 business day.

    Talk to an Expert

    Trusted by

    Other Questions?