Today I basically worked on data, details of the study, and started the module we need to complete to be certified for this study. Since the most efficient way to export data from the app is as one massive dictionary, an email with all the numbers is very difficult to understand. Although we aren’t sure what kind of data analysis we will end up doing with this text, Adam and I both thought that having the data back in dictionary form would be the best way to store it. So I wrote a Python program that takes in one instance of the data set as a text file, and creates a class called DataSet which contains both the actual data (a dictionary whose keys are the indices of the text and the corresponding values are the acceleration at that point) and metadata such as the UID of the user, the name of text (e.g. the first semantical sentence, the third lexical, and so on), and the time it took for them to read. At the moment I also have a function that can create a new text file and put the data into that, but I don’t know how useful this is. I think I could turn that into a JSON file if necessary, once we figure out what we actually want to.
Beyond this, Adam and I spent some time listing all the things we have left to do on this project and started on some of the high priority ones. For instance, we worked out some more details of the study and started designing the survey at the end (it will be here when it is done; it’s mostly empty right now). We still have a lot left to work out, such as:
- What font sizes should be offered
- What sensitivities should be offered
- Which texts should be used
- How many texts can they be expected to read
So I think there is a lot left on that score, which we will probably do tomorrow. I also started the module on ethics in studies and will probably finish it tomorrow.