This dataset contains information on what packages were imported in a sample of 373 R and 4459 Python kernels. Kernels were selected by looking for notebooks that contained the word "import" (for Python) or "library" (for R) but not the default code that Kernels start with. Only public kernels were queried. Then the names of the packages were extracted using the following regular expressions: For R: "[library|require]\\(([a-zA-Z\\.0-9']*)\\)" For Python: "(?<!from)import ([a-zA-Z\\.0-9]*)" This means that only packages that were imported completely were counted: if only part of a package or a single function was imported, it wasn't counted.