Caso de estudio
See what your professor kept coming back to all semester, before you sit the exam
Lo que consiguesNobody can tell you what is on the exam. What you can see is what the course spent its time on. Load the semester's slides, readings, problem sets, and your own notes, then ask which topics recur across all of them, which ones the professor flagged in class, and where your notes are thinnest. You get a ranked list of what the material emphasised, with a link to every lecture and page behind each line.
Two weeks to the exam. Thirteen weeks of lecture slides, a reading list, eight problem sets, and no past papers to practise on. The course had one professor the whole way through, so whatever is on that paper comes out of this material and nowhere else.
The problem is volume, not access. Everything is on the laptop already. What is missing is any way to see which topics the semester actually revolved around, as opposed to which ones felt hard the week they came up.
One semester, one professor, five kinds of material
The signal is spread thin across everything the course produced. A topic that matters shows up in the slides, comes back in a problem set, and gets mentioned again in a later lecture as something you are assumed to know.
- 13 weeks of lecture slides, one deck per week, some of them scanned handouts rather than files
- The reading list, with the chapters actually assigned and the ones only recommended
- 8 problem sets and their solutions, which show what the professor thinks a hard version of a topic looks like
- Your own lecture notes, including the lines where you wrote down that something was called important
- Course announcements and emails, where a syllabus change or a hint about scope usually gets buried
No single one of these tells you what the course emphasised. The pattern only appears when you can see all five at once.
How revision works without any of this
You reread from week one and hope the important parts feel important on the second pass. Time gets spent in the order the material was taught rather than in the order it matters, so week two gets three careful hours and week eleven gets skimmed at midnight.
The other approach is asking around. Someone remembers the professor saying a topic was central, someone else remembers the opposite, and neither of them wrote it down. With no past papers there is nothing to check either version against.
So the topics you revise hardest end up being the ones you happen to remember being told about, which is not the same as the ones the course spent its time on.
Put the whole course in one workspace
This is an afternoon of setup at most, and most of it is waiting.
Getting a semester in
Drop in every folder the course produced
Open "Extract sources" (
Cmd+Shift+I) and drag in the slides, the readings, the problem sets and solutions, your notes, and the announcements. Scanned handouts go through OCR so their text is readable too. Keep them in one workspace: the point is that a topic in week two and the same topic in week eleven land on the same thing.Register the course's own vocabulary
Open "Domain vocabulary" in the left rail and add the terms this course uses. Subjects rename things constantly, and the professor's phrase for a concept is often not the textbook's. Without this, the same idea stays split across three labels and none of them looks important.
Confirm the synonyms it proposes
At "Connecting related items", the app proposes that a term from the slides and a term from the textbook are the same concept. Nothing merges until you confirm it. This is what makes a topic's real weight visible, because mentions that were scattered across three names now count as one.
Ask after the build finishes, not during
Send a message mid-build and a "Knowledge graph still building" dialog offers to wait. Choose to send when extraction finishes. A topic that has not been read yet and a topic that genuinely never came up look identical in a half-built graph, which is the one thing you cannot afford here.
What you ask two weeks out
"Which topics come up in the slides, the problem sets, and the readings all three?"
→ A ranked list. Each topic carries the weeks it appeared in and a link to
every lecture and page it came from.
"How many lectures did the professor spend on each topic?"
→ A table of topics against weeks, so a subject that took three lectures
is visibly not the same size as one that took ten minutes.
"Where did the professor say something was important, examinable, or a common mistake?"
→ Every line from the slides, announcements, and your own notes that flags
emphasis, quoted as written, with the lecture it came from.
"Which topics show up in the problem sets but are thin in my notes?"
→ The gap list. What the course drilled and you did not write down properly.The last question is the useful one and the least obvious. It compares what the course kept practising against what you actually captured, which is exactly where revision time is worth spending.
Every line comes back with the lecture and page behind it, so a claim that a topic dominated the semester can be checked against the slides in a few seconds rather than taken on trust.
What changes
| Before | Now |
|---|---|
| Reread from week one and hope the important parts feel important the second time | Revise in the order the course actually weighted, with the lectures behind each topic linked |
| Rely on someone's memory of the professor calling a topic central | Read every line in the material that flagged something as important, quoted as written |
| The same concept under three different names looks like three small topics | Confirm the names as one concept and its real weight in the semester becomes visible |
| No idea which topics you have under-prepared until the paper is in front of you | See which topics the problem sets drilled and your notes barely cover |
What changes is the order you work in. Reading the material, understanding it, and sitting the exam are still entirely yours.