Delivering high quality data for Tomra

When millions of data points are turned into insightful visualisations, it provides a quality-assured basis for business decision-making.

The need

TOMRA is a major international player which, among other things, produces deposit systems for beverage packaging. Every bottle deposited in a vending machine somewhere in the world leaves behind a lot of data, and if used correctly this huge amount of data contains a lot of useful information. TOMRA lacked a good visualisation tool that could read and process data from both new and older versions of IFS in addition to other older data sources they had.

As the implementation of IFS progressed, it was also seen that there was a need for an integration tool that could synchronise IFS with other internal solutions. Furthermore, there was a need to support business processes across IFS and external solutions at partners and subcontractors.

Finally, TOMRA needed to improve its ability to utilise all collected data from the deposit machines (40-50GB per day of sensor and deposit data) in a way where it was easier to gain insight, a basis for decisions and the possibility of deeper analysis.

Bilde av en mann som snakker med en blond dame utendørs

The Solution

twoday assists TOMRA in several different projects both with consultancy, solution implementation and user training. These are three of the projects in which twoday helps TOMRA:

1. Visualization in Qlik Sense

In IFS, data is collected and produced from many different business processes such as production, stock, purchasing and finance. The challenge was to get a good overview with a business perspective, with flexibility and with a low user threshold.

TOMRA started using QlikView to test and quickly saw what possibilities Qlik could offer. But the user threshold was a little too high in QlikView, so it was only when Qlik Sense was introduced that many user groups in TOMRA could really benefit from the solutions.

When you could also use Qlik Sense as a tool for deeper analysis, closely integrated with the Vertica database (see below), you pretty much covered all visualisation needs in TOMRA with one tool and with one set of skills.

TOMRA valued the ease of use, flexibility and performance that Qlik provided, and they wanted to further exploit these qualities for several business purposes.

Examples of areas where Qlik is used are customer reports on accounting analysis, sales analysis, service analysis, purchasing, supply chain and settlement between business units in TOMRA.

2. Data Integrations with Talend

twoday assists TOMRA with consultancy and delivery of capacity in the field of data integration. The goal of the integrations is to remove manual processes and double registration of data, as well as deliver quality data throughout the organisation. Talend was chosen for this purpose.

Talend is also used for ETL (Extract, Transform and Load) to move large amounts of data from TOMRA's telemetry systems with data from the pawn machines into the column-based analytical database Vertica (see below).

With Talend, TOMRA has increased data quality, reduced manual work, saved resources and gained much better access to its data across systems and business processes.

3. Storage and a quick database Vertica

In spring 2019, TOMRA started a project to define a set of global KPIs. twoday carried out a "proof of concept" on an architecture that TOMRA could use to create good and quality-assured insight into its data from the pawn machines. Based on this prototype, it was decided that the way forward was to use Vertica as a database and 'engine' for analytical functions, Talend to move the large amounts of data and Qlik as a visualisation tool.

Training of Employees

twoday has provided training for both regular users and super users in TOMRA, and the training has taken place both in Norway and other European countries. This has been important for TOMRA to be able to carry out further development of the systems without being dependent on external consultants.

Results

All of these projects/deliveries makes sure that TOMRA has a high quality and good data base to be able to make business-critical decisions based on real data. Data has also been collected from several sources so that many departments get a more correct and faster overview of the business.

Both Vertica and Qlik offer opportunities to work with AI and machine learning. This enables TOMRA to use R and Python for prediction and deeper analysis. This represents a great potential that will be explored further.