Getting students to think carefully about meaningful keywords in their JSTOR searches can be difficult. Constellate's search options provide a great way to help identify potentially useful secondary sources based on specific interests (such as 'words related to baroque in Art History' from the home page here). Below are potential activities you could do with Constellate and its data; feel free to contact Heather Froehlich (hgf5@psu.edu) if you need help implementing these in your classroom.
Let's say you're teaching a class on rhetorical theory and you want your students to identify some particular themes to consider for their final papers. Searching for a specific rhetorical term, such as "logos", on the landing page will provide you several visualizations to use as an introduction to how logos is discussed in a wide range of disciplines. You could narrow these down to more specific disciplines, journals, or time period(s) to see how this term is used in scholarly criticism, and how it might change over time.
Sometimes, faculty want students to be able to identify related literature on a certain topic. For example, the African-American dataset that Constellate has already established would be a great launch point if you wanted students to identify relevant secondary criticism within a well-defined topic. Here, students could choose 2-4 articles from there that could potentially inform a later writing activity. You could create a similar dataset for your students to work on if one doesn't meaningfully exist yet.
Rather than write your own tutorials from scratch, borrow the established Jupyter Notebooks and walk your students through the introductions to text analysis they provide, using a data set derived from a journal or journals relevant to the discipline. Then, you can scale up to more class-specific activities. (Alternatively: if you need to offer some remedial text analysis support, this could be a good resource to direct students to!)