Depopulation Data Challenge
As an example of collective intelligence in action, we established new partnerships that leveraged alternative data sources – such as mobile data, social networks data, job ads data, satellite data – to shine a new light on the complex challenge of depopulation in Serbia.
Why Depopulation Data Challenge?
The complexity of the driving forces behind population dynamics often makes credible data very challenging to come by. Traditional sources of data such as statistical sources (censuses) and administrative sources (visa, residence-work-permit, border data collection systems etc.) have their own strengths and limitations.
Use of non-traditional data sources to better map and understand topics related to population dynamics has been growing in recent years. These alternative data sources are many, and they range from data from public utilities (electricity, water, etc.), to telecommunication operators, satellite imagery providers, and to proprietary data from private companies (job listings, map data, etc.). The value of these large datasets is immense and, although they are not easy to obtain, the ability to combine them with other traditional data could provide deeper insights into various population topics. Big data provides broad coverage that includes some hard to reach populations and often provides information closer to real time.
The Challenge and the Winners
We invited the data science community, academic researchers and everyone interested in the topic of depopulation to propose new data-driven ideas and solutions that can help us better understand what depopulation really means for Serbia. In partnership with UNFPA and supported by GIZ, we announced the Depopulation Data Challenge call in January 2020.
Within two months we received 50 applications from across the ecosystem. Eleven teams made it to the final stage, which included a detailed elaboration of the initial idea. Four winning teams were chosen by a jury composed of representatives of the line ministry, demographers from the Institute of Social Sciences, the Statistical Office of the Republic of Serbia, UNFPA and UNDP.
In addition to financial resources for the implementation of solutions, the teams received mentorship support. Dr Vladimir Nikitović, principal research fellow at the Institute of Social Sciences in Belgrade, assisted the teams by providing domain expertise, guidance on the scientific methods and other technical knowledge necessary for successful project implementation.
Mobile Data
The PopInsight team, made up of researchers from the BioSense Institute in Novi Sad, used mobile, satellite and open data in order to better understand state-level depopulation trends. This team focused on the in-country distribution of population and migration, based on analysis of activity, connectivity and mobility of people in Serbia reflected through telecom data and with data from other sources.
Job Ads Data
The Infostud Data team, from the Subotica-based company Infostud, used data on job advertisements abroad, the structure of jobs offered, as well as the number and characteristics of people from different local communities in Serbia who are looking for employment abroad. By analysing this data, the team explored how many people applied for jobs abroad from year to year, where these jobs were located, as well as how many applications for scholarships and internships abroad were made each year. The solution offers insights on industries that are more and less represented in job advertisements, and on locations to which the Serbian population most often emigrates.
Satellite Data
The Geoanalysts team of researchers from the Geographical Institute “Jovan Cvijić” explored in-country migration, declining population trends in specific areas and population growth in urban areas, by monitoring satellite data, data on night-time lights, power networks and OpenStreetMaps. The goal of this team was to create a permanent networking infrastructure of different data sets that will serve as a unique system for monitoring the depopulation process in Serbia. Insights obtained through analyses of space- and time-sensitive data are presented interactively and provide a much more precise and nuanced overview of the trends and intensities of depopulation and its spatial distribution.
Call to Action
The solutions and data developed within the Depopulation Data Challenge are open and freely available for further experimenting by anyone.
We invite you to use these four platforms and get new insights for your own work or research, to build new products and provide new knowledge from big data, ultimately helping more people.
We hope that these solutions will help you build the capacity to harness data for real-time adjustments and orchestrating new knowledge about depopulation in Serbia.
If this story got you thinking or would like to share something interesting, we would love to hear from you – please drop us a line!
Social Media Data
The Bootstrappers team, made up of researchers from the Harvard University and the Brookings Institution, analysed emigration trends, including industry breakdown of emigrants, by collecting Facebook Ad data. In addition, they explored the connection between the number of emigrants and scientific research output of domestic scientists, using scientific publishing data through the Microsoft Research library – a huge multidisciplinary abstract and citation database of peer-reviewed literature. The team created a website with an interactive map, enabling visitors to see where the people from Serbia are moving to, as well as to filter data and create custom visualizations.