 In the previous videos we talked about text preprocessing, clustering and classification. We worked with green stales, a data set I have prepared in a spreadsheet. But working with spreadsheets and long texts can be a pain. Is there any other way we can import texts into orange? Of course there is. This time I will work with Kennedy's speeches. I have 17 of them in a Kennedy folder, each in its own file. Files can be Word documents, PDFs or plain text files. Here for example, is a speech from Democratic National Convention to load corpus into orange. Open text add-on, place import documents widget on a canvas and open it. Click on the folder icon and select the folder you wish to import. Let us observe our data in a corpus viewer. Here's the speech we've seen earlier. Now we can do some clustering. I'll use preprocessed text, bag of words, distances and hierarchical clustering. I was fast. For details on text preprocessing and clustering, you can check our previous videos. Looks like I have two interesting clusters, one on nuclear arms and the other with Kennedy's presidential addresses. Clustering is fine, but what about classification? Can I tell orange some documents belong to one group and others to the other? Let us put Kennedy's speeches into two folders, say pre-1962 and post-1962. Now reload the folder. Orange recognized subfolders as class categories. If we observe the corpus in a data table, we can see that orange put pre-1962 and post-1962 in the gray class column. You can check our previous videos on text classification to learn how to proceed. Import documents makes it so much easier to organize your files and your research. Today, we've learned how to import our own data for text analysis and how to define class values from scratch.