This is a featured blog post by Dr Belinda Wu, Research Fellow in Development Policy and Planning.
As I pointed out in my previous blogpost, Mind the Gap: The Migration Data, migration has a great impact on any population demographically, socio-economically, spatially and culturally. The result of this complexity is that new tools are required to guide policymakers on a wide range of immigration-related issues, including social cohesion, labour market needs and changes, poverty, inequalities and education. However, if these tools are to have any real impact, migration data first needs a great push to improve.
In addition to the issues discussed in my previous post – such as the inadequacy of current international migration data and the lack of agreement over definitions – which are a result of the way data is collected, processed and the quality assurance that surrounds it, official migration data often lags behind current events. As such, it is simply unable to provide evidence-based support for decision makers during humanitarian emergencies.
Recent studies, however, have recognised the increasing use of social networking platforms amongst refugees to seek help and to express the hardship and difficulties they face during their journeys to their new destinations. In turn, this creates a powerful opportunity to study these sudden large-scale international migrations using big data methods; methods which have also been successfully applied in modelling various human behaviours. If the patterns of the migrations concerned can be understood correctly and accurately predicted using the almost real-time data available from online social networking platforms, it will allow the local governments and aid agencies of the countries concerned to be considerably better prepared. As a result, they will also be better able to help the refugees and mitigate the impact of the problems they meet both during their journey and in the migration camps. Such studies can therefore be enormously helpful in ensuring the migrants’ safety and wellbeing, as well as in facilitating better governance of these large-scale human movements and their consequent developmental issues.
One recent cause of migration has, of course, been the war in Syria, following which the migration of millions of people has led to complex humanitarian, social and economic issues; not only for the refugees themselves, but also for the hosting countries and the various governmental and aid agencies involved. In one such study in which I took part, therefore, a group of OU academics (Dr Patrick Wong, myself and Dr Soraya Kouadri), Dr Haiming Liu (from the University of Bedford) and consultant Dr Smarti Reel explored the usefulness of Twitter data in studying the Syrian refugee migration patterns.
Although, on the plus side, it is often freely available and can provide almost real-time information, big data gathered directly from the internet is also messy, lacking the standard of official datasets. Also, volume is only one aspect of it. Other attributes include variety, velocity, value, and complexity. As a result, considerable skills and resources are needed to make use of big data.
Our pilot study therefore developed a framework to retrieve, filter, analyse and classify the Twitter data. The first issue we needed to tackle was to identify prospective refugees and to understand their use of Twitter, as many tweets concerning refugees are actually posted by humanitarian workers, activists, academics or journalists. As such, the data related to their tweets and tweeting behaviours will be very different to that of the refugees themselves. This issue led to our developing an intelligent classifier able to determine automatically those tweets that came from potential Syrian refugees. Testing the classifier with samples of historical Twitter data produced promising result, with a 62% correct classification rate.
This preliminary study thus clearly demonstrates the potential of using social media data to identify refugees and their movements. As a result, we plan to continue in our efforts to achieve our research goal of providing in-time predictions of migration patterns to support decision makers.
As part of this process, I am convening a panel at the DSA (Development Studies Association) Conference, Global Inequalities, in June and papers are invited on the theme: ‘New’ Methods in Research and Communication of Global Inequalities. You can read the full invitation for submission of papers here, but please note the deadline is 9 March.
Dr Belinda Wu joined The Open University in January 2016 as a Research Fellow in Development Policy and Planning. Upon completion of a PhD in Human Geography with the University of Leeds in 2013, she became the named researcher for the ESRC project SYLLS to generate a synthetic microdata for UK Longitudinal Study in the Centre for Advanced Spatial Analysis, UCL, where she also used a Complexity approach to study the International Development Aid for the EPSRC project Enfolding 2014-15.
Previously she has worked on two ESRC projects: Moses and Genesis, at the University of Leeds. Both ambitiously attempt to provide a generic modelling and simulation basis for various decision making support and e-Social Science research. She is also a Fellow of Royal Geographical Society (FRGS) and a Member of the Institute of Logistics and Transportation (MILT).