Tracking migration with Facebook
19 Feb 2019 10:59 AM
Social media can provide valuable real-time data on migration, with new research using datasets from Facebook's advertising platform to uncover trends and migrant numbers.
With the ongoing debate around immigration, 'hard' research evidence on migrant numbers and trends is crucial – but there are still huge limitations on available data for measuring migration. There is no consistency in the definition or quality of European migration data, with countries separately collecting data using different systems and designs.
"Especially for migration there is a lack of quality and quantity in data. Social media can give a much faster picture of migration, as we can get data right up to the minute," explains Francesco Rampazzo, PhD student at the ESRC Centre for Population Change and the Max Planck Institute for Demographic Research. He leads an ongoing research project using Facebook data to provide a clearer picture of the numbers of migrants in the UK.
When it comes to robust and reliable migration data, population surveys and census data are the safe and obvious sources – but these are 'snapshots' of the situation at a particular point. Social media, however, can provide an ongoing data stream. Although Facebook doesn't reflect the whole population of a country – and therefore isn't a representative source of data – it's attractive as a data source as it's updated in real time, with several ways of filtering information.
Rampazzo is focusing on anonymised grouped data from Facebook to estimate the amount of people migrating from other European countries to the UK, based on the change in the location that they specify. The aim is to create the first longitudinal geolocated dataset from Facebook’s advertising platform (which provides data to advertisers), through a weekly data collection that started in December 2017 and carries on till the end of his PhD (October 2020).
Using so-called 'digital trace data' from social media can potentially be a gold mine for demographic research in sheer quantity and detail; the Facebook ad platform has even been called 'a new digital census' by researchers. "I think it's something that is becoming more accepted in academia," Rampazzo says. "I believe it's pioneering in this research to use Google search, Twitter and other internet data to estimate population trends. But it's still in its infant stage – it's becoming more accepted, but we have to dig into the quality of the data."
Assessing the quality of the data from social media can be challenging. How does the data stack up against traditional census data in terms of bias? Is it collected or processed in a sufficiently similar way to make it possible to compare with census data?
"We get biased information from social media, but there's also bias in traditional sources. So we need try to understand and identify data bias in social media, and then complement with traditional data where the bias is already known and can be taken into account," says Rampazzo.
Early findings suggest that the Facebook data can provide a valid picture of migration trends. "Our preliminary results from Facebook are promising, showing proportions close to estimates from the Office for National Statistics. We believe that this is an encouraging and new direction for future work – not just in the UK, but at a global level."
Further information