PredictHQ Beam

The automated correlation engine to accurately reveal the events that drive demand for your business.

Beam turns 72 days of work into 52 seconds

Combine Data Sources Input

1 Impute Values
+ Check Quality 0.06 sec vs 2.5 days 2 Time Series Decomposition 48 sec vs 45 days 3 Anomaly Detection 2 sec vs 7 days 4 False-positive Detection 1.5 sec vs 7 days 5 Correlate
Demand 0.5 sec vs 11 days

Export + Integrate Event Data Output

The input

Combine unstructured event data + transactional data

PredictHQ converts unstructured, dynamic event data into a workable dataset, allowing data science teams to use intelligent event data in machine learning models. PredictHQ Beam combines and correlates our event data with your transactional demand data.

Impute Values + Check Quality

Beam automatically detects missing values from your transactional data and imputes values using our advanced time series reconstruction technologies. It then systematically checks data quality with stationary attributes and rolling standard deviation.

Time Series Decomposition

We decompose your time series transactional data using IteSSA nonparametric machine learning algorithms to automatically detect weekly and monthly cyclicities, seasonal patterns, and long term and short term trends.

Anomaly Detection

Using the patterns identified in step two, Beam automatically detects the incremental anomalies in your transactional data based on the re-constructed demand patterns.

False-positive Detection

False positive detection algorithms detect the positive anomalies which are not generated by attendance-based events and then adjust the incremental demand estimation.

Determine Incremental Demand + Correlate To Events

PredictHQ uses FDR (false discovery rate) to calculate the strength level of the incremental demand. We segment all the detected incremental demands into strong, medium, weak and no impacts. We join the detected incremental demands with attendance-based events impacts as well as holiday events counts to correlate the time series transactional data with events.

The Output

Export the results and seamlessly integrate them into your prediction models

Download the full results to understand which events impact your demand and use the Features API to improve your forecasting models.

Demand graph

High-impact events list

Custom PDF report

Easily visualize and scroll over your demand signals and view corresponding event impact in Control Center.

Date	Event Impact	Example events driving demand
18 Mar 2020	34,396	44 PHQ 89 Local 92 Aviation North American Society for Trenchless Technology
29 Mar 2020	20,373	75 PHQ 92 Local 41 Aviation NBA — Dallas Maveriks vs New York Knicks
30 Mar 2020	14,373	65 PHQ 71 Local 33 Aviation Seattle Mariners vs Texas Rangers
1 Apr 2020	14,023	86 PHQ 80 Local 39 Aviation Minnesota Twins vs Texas Rangers
4 Apr 2020	29,960	44 PHQ 89 Local 92 Aviation Severe Delays — Toupo Airport (TUO)
5 Apr 2020	8,771	75 PHQ 92 Local 41 Aviation Queen + Adam Lambert
10 Apr 2020	6,109	65 PHQ 71 Local 33 Aviation New Zealand vs India
16 Apr 2020	25,667	86 PHQ 80 Local 39 Aviation Adele Concert
22 Apr 2020	18,360	44 PHQ 89 Local 92 Aviation New Zealand vs South Africa

Scroll through the list of events and their impact. Plus, you can click into the days you want to drill into to understand what individual events make up the demand.

Download a customized report with the data broken out in a variety of ways, including in graphs, lists of events and impact by category.

Start using Beam today

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