Analytics

Advanced, model-based data processing engine at your finger tips

PhixFlow analytics are built using drag and drop models and spreadsheet-like expressions. Objects themselves are highly configurable allowing PhixFlow to handle both simple and very complex analytics.

PhixFlow Platform analytics diagram

Scheduled or user driven

Analysis models are configured for each piece of data processing that you want to carry out on a regular basis. The models can be set up to be run on a scheduled basis e.g. hourly checks of order progress or can be linked to menus and buttons on configured user screens. This allows models to be run at any time the user requests and in the context of the data that the user is looking at, e.g. to trigger an update to one of your systems based on a user request, or to carry out more in-depth analysis of a highlighted problem.


Performance

If your models need to support the processing of large numbers of records then all the performance features we’ve tuned over the years to process massive volumes of data will benefit you. Growing out of the telecoms industry where it is not unusual to process hundreds of millions of call records per hour, our analytics engine is massively scalable.

PhixFlow analytics is “multi-threaded” i.e. each “record” of your data is processed in parallel enabling processing rates of hundreds of millions of records an hour. PhixFlow also manages intelligent in-memory caching of reference data to reduce the need to query databases too often and provides full control on the number of parallel threads used to read and write data from your systems.


Enriching and blending data

At each analytics step, PhixFlow objects allow you to:

  • read data from any source.
  • carry out multiple calculations on each data attribute using spreadsheet-like expressions.
  • enrich the data with lookups to internal or external reference data.
  • merge data records from multiple sources e.g. a single step could merge 4 or more data sources.
  • de-duplicate data according to the rules that you define.
  • filter & sort e.g. to just pass through exceptions for processing.
  • write data to any target e.g. databases, spreadsheets, files, emails, web services.

Some analytics objects

 

database
Database

Specify how to connect to remote databases with user and password details.

 

merge
Merge

Blend data from multiple sources. E.g. 4 sources merged in a single step.

 

calc
Calculate

Perform a set of calculations on every row of data.

grouped-calc
Grouped calculate

Transform data with reference to data in a grouped set. 

aggregate
Aggregate

Aggregate data over grouped sets.

fx
Expressions

Spreadsheet-like expressions used to transform data. 

macro
Macro

Re-usable expressions contained in a library managed by you.

filter
Filter

Filter expressions to select which data flows between modelling steps. 

spreadsheet
Spreadsheet

Read or write to spreadsheets. 

 

web-services
Web service

Integrate with web services to read data or trigger actions on remote systems.

 

db-writer
Database write

Specify what to write to a database based on data flowing through PhixFlow.

 

email-sender
Email sender

Define the details of emails to send with optional attachments and dynamic content.

 

Dig deeper

PhixFlow solutions put you in control by letting you design applications to fit your processes.

Our solutions