A collaboration rooted in innovation
Founded in 1999, IPKO Telecommunications began as Kosovo’s first nationwide internet service provider. Since then, it has grown into one of Europe’s fastest-growing telecom companies, known for its focus on innovation, service quality, and customer satisfaction.
As the company evolved, and recognising the value of its data, IPKO saw an opportunity to support government decision-making through data-driven services. This ambition led IPKO to seek out the right partner who could help turn that vision into reality.
In June 2024, IPKO reached out to us for our expertise and tools to help unlock the power of data for public good. Together, we began shaping a project aimed at delivering smarter, more informed governance for Kosovo.
Having identified infrastructure and transport as sectors of priorities, IPKO asked us to demonstrate their data value for the Kosovo Government to stimulate demand and investment in utilisation for these two sectors.
Our solution
We focused on enabling and supporting IPKO’s team with algorithms, troubleshooting and organisational guidance to produce aggregated data from pseudonymised call detail records for infrastructure and transport applications. From these aggregated data, we produced visualisations and analysis, bringing IPKO staff along each step of the process.
With this approach, we worked on building IPKO’s technical capacity by teaching them how to use our code and produce aggregates, and strengthening their analytical skills by teaching them how to interpret the findings as we analysed the data.
Understanding needs and environment
Our engagement started by understanding the types of applications IPKO wanted to support. We worked hand-in-hand with their team to discuss potential use cases and solutions based on their data.
Understanding IPKO data has been instrumental in those discussions. From data warehouse set up, hardware infrastructure, to dataset structure, data types, retention, and cell tower coverage, we had to understand the data infrastructure to determine which solutions could be developed. From these, we were able to adapt our code to run in their systems, and determine the types of analysis we would conduct from the produced aggregates.
The purpose of the collaboration was to enable IPKO to understand how to extract value from their datasets, aligning with IPKO’s investment in L&D for their staff.
With this in mind, our collaboration primarily focused on training IPKO staff to produce aggregates from our code. The training programme included:
- An introduction to our work, the wider Flowminder machine and our most refined solutions
- Technical data training, explaining the aggregates we’ve asked them to compute and providing interpretation of what these aggregates actually mean. This session covered how aggregates are turned into indicators or more developed estimates, including the difference between commercially sensitive aggregates and sharable indicators, and quality assurance principles and methods.
- Analytical session indicators / analysis, with some examples of the analysis process / live coding session. In these sessions, we went over the actual work done to provide the final outputs for dissemination, with a stream focussing on interpretation and another on data visualisation.
Blerim Tahiraj, with whom we collaborated on this project, shared his insight on the value of the work on LinkedIn - read more below
Photo credit: Blerim Tahiraj (LinkedIn)
