This analysis job takes a single dataset and plots it by publication date.
To start this job, select the question: “When were a given set of articles published?”
This job results in a graph of the publication dates of your dataset, as well as the counts downloadable as a CSV file. When the graph is computed, you may optionally choose to divide the absolute counts of articles published in a given year by the counts in another dataset or in the entire corpus, giving you the fraction of articles in the larger set which appear in your search (see the options section below for more information). By simply counting the publication frequency over time, you can actually answer a wide variety of questions:
How much of the literature in a given time period was focused on a given topic? (Input: a dataset created from a search for this topic, normalizing against the entire corpus)
When you run this analysis, by default it will simply return the number of articles published in each year in your dataset. Often, however, you would like to know these counts as fractions – for instance, how many of the articles published in each year in the entire corpus are found in this dataset? We call this normalizing the counts.
To normalize, check the box and choose the dataset (or the entire corpus) that you’d like to normalize by. This will take the counts of articles per year in this dataset and divide them by the counts of articles per year in the dataset you select.