Trends in Economics Papers
2-grams are supported, e.g. general equilibrium.
JEL codes are supported, e.g. D, D4, or D43.
Singular and plural nouns, etc. are treated distinctly.
Stop words and near-zero trends are omitted.
While developing and testing this page, I obviously did some searches myself. Here are some silly and some serious examples.
Last but not least, COVID-19.
The trends are compiled from titles and abstracts of economics papers, as found in the distributed RePEc dataset. For downloading the dataset, I use a collection of Python scripts that I wrote myself, these can be found on github. The RePEc dataset contains some non-English papers but also many papers which I would not classify as papers in Economics. I therefore filter on papers that are in English and that have JEL codes. In the last dataset update, there were 563,854 such papers during the past 30 years. Text processing is minimal. First, stop words are removed. Second, strings of consecutive Latin characters are extracted and treated as words. Third, for every word the number of papers is counted where that word occurs in the title or in the abstract at least once. The RePEc dataset has grown substantially over the years, and the numbers for earlier years are less informative than the numbers for later years.