This paper presents a unique tax revenue dataset for Sub-Saharan Africa, which main innovation is the level of detail it provides about tax revenue sources for a large number of countries (41) and over a long time period (1980-2010). The paper describes how the dataset was constructed, identifying along the way problem areas in tax revenue statistics in Sub-Saharan Africa and possible improvements. A graphical analysis highlights revenue performance over time and across three dimensions: income levels, the relative importance of tax revenue from extractive industries, and trading groups (free-trade areas and customs unions). The dataset, available at www.ferdi.fr, should be useful to a wide range of users and researchers, including academics, tax policy practitioners and advisers, and revenue and customs administrations.