"The increasing use of the internet and smartphones in developing countries has a big potential for early warning systems"

Regine Weber, ZEF junior researcher, reports about her doctoral research on innovative early warning systems on food security.

What is your research topic?

My research deals with innovative approaches to food security monitoring in developing countries. In my thesis, I analyze existing early warning and food security monitoring systems and their indicators, to develop a theoretical framework for monitoring risks. I further explore the potential of Big Data for near real-time monitoring in developing countries and test, in a case study in Kenya, whether Google Search Queries are a valid source for food price tracking. In addition, I am currently implementing a new approach to the monitoring of risks to food security in cooperation with Welthungerhilfe, a large German NGO. We are piloting an innovative data collection initiative in Kenya, Afghanistan and Mali. This pilot study is SMS based and gathers information on the local food security situation that is highly geographically disaggregated. It is based on a network of partners and practitioners with local expertise.

What were the reasons to pick this topic for your doctoral thesis?

The increasing use of the internet and smartphones in developing countries contributes to the generation of fast and in-expensive information that comes from affected people themselves. This has a big potential for the development of faster, more precise and more comprehensive monitoring and early warning systems, as well as for better humanitarian response strategies and ultimately, to achieve better food security outcomes. Since my undergraduate studies, I have been interested in understanding the drivers behind food crises and also the mechanisms to reduce their extreme impact. While exploring topics for my doctoral thesis, I was quickly fascinated and captivated by incredibly interesting and innovative new initiatives and opportunities that were emerging in the realm of food security monitoring – and it was at this moment when I decided that I would like to contribute to this topic with my doctoral research.

What are the major challenges that you face during your research?

Like every PhD student, I have been dealing with a variety of challenges. One challenge is related to accessing information that is online generated and, theoretically, publicly available. Much of the user generated content is, however, still owned by the private sector and, unfortunately, many details are not disclosed. Twitter, for example, won’t let you download the full Tweet history and Google will only provide a percentage share of Google search queries in relation to the non-disclosed total amount of search queries – which has strong implications for a potential analysis. Furthermore, this data still lacks geo-detail and is to a large extent aggregated at the country level – which is unfortunate for early warning and near real-time monitoring, as information is required not only to be fast but also to be geographically disaggregated. Only information with a high geo-detail allows us to pinpoint localized risks and to identify pockets of food insecurity.

Since the usage of big data to monitor the food security situation is quite a new idea there must be room for improvement. What are the possibilities, what are the limits?

Food security monitoring in developing countries is currently at an innovative stage – but the full potential of increasing smartphone and internet adoption rates is yet to unfold. This information revolution holds vast potential, way beyond research, for the democratization of information and for bottom-up movements. Limitations, however, stem from the fact that, as mentioned before, much of the data is held by the private sector and access is limited. Another point is that Big Data brings about ethical concerns that are, amongst other things, related to data privacy, ownership and security. One could, for example, think about whether agreeing to a vast list of Terms and Conditions when using Google is informed and fair consent. The ethical discourse in that regard has not evolved in symmetry with the technical possibilities. This topic, as well as its implications, urgently needs to enter into the broader societal discussion.

To work with big data only requires a computer but your doctoral project requires also a lot of travel. Even Afghanistan might be on the itinerary list. Why is that?

Data collection initiatives that do not require enumerators to be physically present and move from household to household are particularly interesting and hold big potential in regions where movement is limited and too risky, due to ongoing complex emergencies or conflicts and where official data collection activities might have broken down or have been suspended. Nevertheless, piloting such an initiative requires close cooperation and network and trust building with local institutions and partner organizations. This is the only way to ensure adaption to the local context and reality and to understand which initiatives are in place and where can we add to.

Foodmonitor.org is a monitoring system which you helped to develop at ZEF. How does it work?

That’s right. Food Monitor currently has four indicators that monitor different aspects of agricultural markets on a global and national scale. These indicators are

·         Price Transmission Indicator: monitors the transmission of international price movements to markets in developing countries;

·         Food Security News Hotspots: a count of news stories that are being published in relation to agricultural markets;

·         Global Supply Indicator: estimates whether the global supply situation matches global demand;

·         Excessive Price Volatility Indicator: shows abnormally high price swings on international markets.

Food Monitor’s innovation lies in its web scraping algorithm, which automatically collects and stores data from various websites, recalculates this data and then updates results on a daily basis. This sophisticated mechanism provides the opportunity to produce close to real-time information that can then be shared easily. A key feature of the website is the integration of Twitter – the Food Monitor account has been specially integrated to ensure any changes in the levels of risk are communicated immediately via an automatic tweet.

The launch of Food Monitor was an exciting time for the team but the work does not stop here. The idea is to extend Food Monitor in the future through the integration of local food security information on Kenya, Afghanistan and Mali. Further, our challenge now is to keep users engaged and interested in the website in order for it to fulfil its full potential. At the moment, Food Monitor’s exposure continues to grow thanks to our social media activities.