Claude as research assistant: How to read scans efficiëntly?
I'm a PhD researcher and all questionnaires were filled in on paper (due to the setting), leaving me with an extensive amount of manual work (data-input into excel & spss). I tested if Claud could read my PDF scans (+- 35 pages per questionnaire) and transferred them into an excel based on my codebook. It turned out to work quite well, with some mistakes here and there (but I ofcourse check everything). He could process about 10 scans per session in the beginning, however now it stopped working and he gets immediately to my usage limit (I have a pro subscription). I used to let it run in the background and then after a few days when a batch of +- 30 scans were ready, I checked everything and so on. Now, that is not possible anymore, which is quite annoying. I tried to open another conversation, to lower the scan quality and asked tips to claude himself but nothing really seems to work. Anyone that has experience with this and has some advice? Maybe other AI tools are better for these kind of tasks?