Embracing the Principles for Digital Development: Leveraging Big Data Sources for Policy Evaluation in Crisis Settings
COVID-19 has exacerbated challenges facing vulnerable populations world-wide. The lack of adequate data for informed policy-making in crisis settings pose a particular challenge in Lebanon and Jordan..
Data-Pop Alliance (DPA) and the United Nations Economic and Social Commission of Western Asia (UN ESCWA), partnered with the Central Administration of Statistics (CAS) of Lebanon and the Department of Statistics (DoS) of Jordan, to close the data gap which has hindered decision-making processes especially under crises pressure. The project, “Leveraging Big Data sources for Policy Evaluation and Analysis in Crisis Settings”, created a web platform from which policymakers can draw insights and recommendations based on policy impact to assist crisis responses, similar to COVID-19, in a more agile and targeted manner.
This project builds upon a prior collaboration, (in line with the Reuse and Improve Principle) between DPA and UN ESCWA. It explores the potential of non-traditional data sources for policy creation, planning, and assessment for improving the living conditions of Syrian refugees and their host communities in Lebanon. The project report can be found here.
Using the UN ESCWA developed observatory for COVID stimulus and policy tracker, we reviewed, ranked and selected five policies in each country. The selected policies focused on the major events taking place in both countries. For Lebanon, they included: social solidarity to address the repercussions of COVID-19; hybrid education; Banque du Liban intermediate circular to grant exceptional loans; solid waste management following the Beirut port blast; and the COVID-19 management plan to increase capacity of health centers. As for Jordan, they included: social solidarity; hybrid education; emergency cash transfer to support the most vulnerable; and protection and response measures for persons with disabilities.
The Use of Open Data Sources
As a data-driven project, the main goal was to study the usability of different non-traditional, near real-time and open data sources with the potential to produce applicable findings without undergoing extensive and expensive data collection processes. We conducted an analysis of the data ecosystem to better understand which data sources could yield better results. In the end, the sources used were: Twitter, Facebook, GDELT, Google Trends, and Google Mobility. Throughout the analysis of this data, we were able to learn about the development and relevant changes in sentiments, opinions, and demographic characteristics of those impacted by the policies.
In Twitter’s case data was collected by using keyword-based queries in Python to scrape public tweets associated with topics of interest (e.g. COVID-19 related policies), both in English and Arabic. The scraped keywords were analyzed to extract the sentiment trend(s). GDELT, though not a social media platform, is an open dataset built from monitoring the world’s broadcast and web news. Just as with Twitter, we scraped the articles alluding to the aforementioned policies to study the populations’ sentiment variation.
Using non-traditional sources allowed us to understand how diverse COVID-related measures were perceived during the first months of the lockdowns in Lebanon and Jordan. However, these sources have drawbacks too. Technical and scientific, institutional and cultural, legal and ethical, commercial, and financial challenges are often encountered. For example, while companies such as Facebook and Google share data during identified crises, they are not always available for all locations or categories, or contain sufficient geographical granularity.
A Platform for Effective Policy Making
As mentioned above, the ‘UN ESCWA – DPA Big Data Insights for Public Policy Decision-making’ platform was the ultimate outcome of the project. Though still in its final development phase, the dashboard uses Big Data-derived tools to draw policy-making insights.
The platform was designed for and with the end users: civil servants, ministries or other agencies tasked with risk mitigation and response, and development practitioners collaborating with both governments. A capacity building workshop was conducted with UN ESCWA, DOS and CAS to introduce the dashboard capabilities and explain its main guidelines to all interested National Statistical Offices in the Arab countries. Moreover, considering the sensitive use of this platform, preventive measures were taken to address privacy and security concerns.