…
Intelek Data Pipes

Data API software

Access your data and run your algorithms via API, from any computer system

Intelek Data API facilitates access and interaction with the data in your Data Warehouse by any program without the need to grant direct access to the database. Moreover, it makes it possible for computer systems on which it is not easy to build artificial intelligence solutions to remotely execute this type of algorithm via API and work with the results in real time.

Query via API any table/data from the datawarehouse
Execution via API in real time and on demand of algorithms developed in Python or R
Governed and inheriting access permissions
Monitoring the use and quality of execution of algorithms.

01Data query via API

For this reason, at Prenomics we have created an API that allows any application to communicate securely with the data warehouse. In particular, the Intelek Data API allows:

  • Make queries to data warehouse entities without needing direct access to the database. But instead, making authenticated requests.

  • High level of flexibility in these queries, in which it is possible to select variables, filter data, group it and perform simple transformations.


When you have rich data in a data warehouse, you often want to use it within the scope of some company IT system. However, this step is not always easy without compromising the security of multiple systems that are hosted in different environments.

02On-demand execution of analytical algorithms via API

From Prenomics, we have enabled a process execution API in Python and R that facilitates the rapid uploading of this type of process to production and programmatic execution by any computer system.

  • Execution of Python and R processes from internal or external data.

  • Monitoring of the execution quality of the algorithms.


It is common to want a corporate system to execute analytical processes (eg, obtaining the most relevant product to recommend to a client). However, corporate systems are not always prepared to build this type of algorithms in a simple way.

03Illustrative case

E-commerce that wants to incorporate the functionality of displaying products dynamically according to the interests of the visitor to its store. However, developing this intelligent recommendation engine is not at all easy since the ecommerce software used by the company and the prefabricated solutions that exist in the market do not have the expected behavior.

Solution

  • In a first phase, a customer segmentation algorithm is built and these segments are dumped into the data warehouse. This algorithm is run programmatically 1 time per day every morning.

  • In a second phase, and this is where the Intelek Data API comes in, a recommender algorithm is built from customer data, its segment and its recent browsing history. This second algorithm is executed in real time during the client's navigation. To do this, the ecommerce software makes calls to the Intelek Data API passing the customer or visitor identifier and it returns the list of products to be displayed suggested by the recommender algorithm.

Impact

  • Being able to develop the recommendation algorithm that will work in production from R or Python regardless of the ecommerce technology.

  • Quick upload to production of the algorithm and its changes without affecting the ecommerce software.

  • Enable the benefits of this type of algorithms without the need to spend money on infrastructure and its maintenance.

Related content