Food and Agriculture Organization of the United Nations
    FAO Data Lab

    Analysis of (social) media

    The Data Lab extracts information from the tweets posted by the accounts of more than 270 newspapers worldwide, and from the news retrieved with complex queries executed on Google News.

    The resulting text is treated through standard text mining techniques and automatic or semi-automatic strategies.

    The final products are:


    Covid-19 impactCovid-19 impact analyses: an open-access tool developed by the DataLab that gathers, organises and analyses daily information on the impact of the COVID-19 pandemic on food and agriculture, value chains, food prices, food security and undertaken measures. Its ultimate aim is to provide countries with facts and information on how the pandemic is impacting on food value chains to support their decision-making. 
    social unrestA social unrest analysis tool: a platform that visualizes social mood trends for each country, identified through the analysis of the tweets published by newspapers' accounts from 150 countries. Such tweets were published in 4 languages over a time period that goes from January 2020 to the present day.
    News searchA semantic search engine for all the news collected by Google: a searchable platform of news items related to the impact of COVID-19 on food value chains (automatic extraction from the Google News engine, with tailored queries). Results can be filtered by language (English, French, Spanish, Russian, Arabic, Chinese), commodities, countries and topic (food chains, socio-economic, Covid-19, civil unrest, food losses and wastes...).
    TweetsA user-search engine for the latest tweets: the DataLab created a system that extracts the latest tweets by considering one or more search keywords; the resulting texts (including the ones that are linked in each tweet) are then analysed automatically to identify the main topics and the terms that are most representative for them. 
    SelectionA search engine that acts on a selection of previously collected tweets/news: in this case the tweets/news collected by the DataLab are filtered by means of one or more search keywords; the resulting texts are analysed automatically to identify the main topics and the terms that are most representative for them.