How to run a succesful data collection campaign

Published on: September 3, 2019

High-quality data about company assets is essential if you want to do efficient preventive maintenance. However, for many asset managers, setting up a data collection campaign may seem like an insurmountable task. How do you organize such a comprehensive project? How can you be sure that you collect all the data you need? And can you trust the data you collect? Kapernikov has been addressing those questions for many years. Based on our work with large utility and infrastructure companies, we can offer a solid framework for data collection campaigns.

Kapernikov has been running a data collection campaign for Infrabel, Belgium’s railway network operator, for more than three years. And a new data collection campaign is in the pipeline. This has allowed the company to build up a lot of data collection expertise.

“The Infrabel data collection campaign has been a huge task for Kapernikov, but a critical one for the maintenance and safety of Belgium’s railway network,” says Hans Nickisch, data consultant for the Infrabel project. “The project has been invaluable for our data collection expertise. And although every data collection project has its specific challenges, our experience and methodology are a basis for future projects.”

Efficient data collection methodology

Data collection is a cyclical process that consists of different steps. Let’s have a look at the data collection method we have used for Infrabel’s railway assets.

Step 1: Collecting readily available data

Many companies state that people are their greatest assets. This is certainly true for data collection. Company knowledge is not always documented, yet it is present with in-house maintenance workers and employees. Therefore, interviewing the Infrabel maintenance team about the state of the company’s assets was a first step. Not only did the acquired information allow us to simplify the planning, but by asking the team’s input, we also gained their goodwill.

Step 2: Day-to-day work planning  

Based on the first internal round of data collection, we were able to set up a rough day-to-day work planning and project forecast for the actual on-site data collection.

Step 3: Collecting data with the right tools

Site surveys and data collection can only be efficient when you use the right tools. Kapernikov provided the Infrabel data collection team with all the technical tools they needed. Operational assets on site were labeled, scanned, photographed, and documented through a thoughtfully developed script, so the operator was unable to miss anything.

Step 4: Processing collected data

All collected data was transferred and synchronized in the cloud. At this stage, the Kapernikov team stepped in again for the evaluation of the data quality. All the data was manually checked, cleaned, corrected and updated.

Step 5: Automatic data processing

Data processing is a continuous interaction between manual and automated data processing. Both complete each other. Kapernikov’s complex analytics can be used to mine for data patterns and to make useful conversions. From simple conversions, like replacing a number, to more complex operations, such as retrieving station names, our algorithms cover a range of possibilities to make customer data more valuable.

Step 6: Adapting the planning

After the first data collection and processing rounds we were able to better assess the amount of work and effort we needed to put into the remaining data collection. This allowed us to update the rough planning into a more refined schedule.

Data collection campaigns

“Kapernikov helped Infrabel to make the data collection process more uniform and also provided the appropriate tools for data collection,” says Fabrice Dumont, data consultant for Infrabel. “This has been the added value of Kapernikov. The systematic approach for data collection is the ideal basis for more efficient preventive maintenance and ultimately for safer infrastructure along Belgium’s railways.”  

In the case of Infrabel, the data collection was planned as a continuous flow of activities over the course of several months. The Infrabel data collection campaign included more than 150.000 checks of operational assets. By batching the work in a continuous effort, Infrabel was able to perform the data collection in a much faster and more cost-efficient way.

“Three years of data collection have taught us a lot about planning and managing a data collection campaign,” says Vincent Goffin, data consultant at Kapernikov. “We now have a data collection framework that makes sure all data is relevant and correct, and that allows us to react proactively to possible challenges that lay ahead.”