INFOTEL THANKS YOU FOR YOUR APPLICATION

A recruitment consultant will contact you soon.

Banque de France – setting up a web portal

Promote innovative uses of data and accelerate the production of information and indicators by creating a portal for sharing and searching economic information

Creating an open database of data

The Banque de France handles large amounts of statistics from different departments (statistics, finance, prudential supervision and excessive debt). These come from numerous sources and are stored in separate silos with different formats.

Background

Since November 2013, Infotel has been working on the MUSE project at the Banque de France to set up an in-house ‘open-database’ of statistical data.

The aim was to set up a shared portal and seek economic information to promote innovative uses of data and speed up the production of information and indicators.

Solution

The project was initiated following an analysis of the existing data and digital reproduction requirements.

During the PoC, two preliminary stages made it possible to select the technologies capable of meeting Banque de France’s needs before proposing ways of better using statistical data owned by the Banque de France.

Due to the large volume of data involved, and in order to provide comprehensive storage of the information available (while also providing sufficient flexibility to interact with the data), a data lake was created to host the information, homogenise the formats, develop the research capabilities and manage security from an architectural perspective.

Results

The solution chosen meets the Banque de France’s requirements and makes it possible to collect and aggregate data from different organisations, forecast and detect trends, offer a 360° view of the client and detect fraud and suspicious behaviour.

The comprehensive solution offers free search, by facet, dynamic data statistics, and an opening in the form of Web Services plus a preliminary aggregation in Hadoop.

Testimonials

The expert's viewpoint

Since 2013, MUSES has been a groundbreaking project in the field of Big Data at the Banque de France. We are currently building on what has already been accomplished and are also working on redesigning applications around new web technologies such as Angular 4+. Everyone can play their role and contribute to homogenising the range of applications drawing upon the same technologies and following the same graphic guidelines.

Applications expert
Clément LABARRIERE

Method and resources

Team and timeline

  • Date of the study/PoC: 2013
  • Length of the operation: 6 months, 300 j/h
  • Speakers: dedicated team of three Big Data experts
  • 2015 and 2016 project – 3 half-time experts, 3 developer-designers
  • Maintenance started in January 2017

Deliverables

  • Feasibility study
  • PoC
  • Projects underway in the  NTIC Banque de France service centre – 3 FTEs over 2 years
  • 2016 – 2017 developments: Switch to SparkR

Technological environments

  • Technological choices following Benchmark: ElasticSearch and Angular JS, HBASE and Hadoop
  • Industrialisation of the cluster, increase in data volume (25To)
  • Agile development
  • Integration of security issues
  • Additional dashboards with Kibana