Okay. Yes. This sounds like the dumbest question.
I just have a PDF that has a bunch of data in the ugliest possible way and I’ve attempted to pull it into an Excel sheet by myself and mess with the powershell to make it better, but it’s awful, I’m too novice, and I frankly don’t have the time.
This is a personal request, not actually condoned by my workplace, but the data wouldn’t be sensitive. So could I find someone on like fiverr? Or, I dunno where to even look honestly.
Any suggestions?
The obvious answer here is to get on an excel community, post your attempt at it and claim it is working perfectly, then wait for someone more knowledgeable to angrily correct you.
Something that may help, if you do end up doing this manually, is utilizing alt+click to highlight in the pdf. It’ll select in a box shape instead of reaching the end of each line like a paragraph. Each line will then get pasted as separate rows in excel. It doesn’t split into columns, so you still have to go with individual vertical selections. It also will break apart multi-line entries like I saw in the description field on your sample, so this will have to be highlighted individually normally. But at least it adds a little speed to those triple coordinate items. Maybe.
You can parse this kind of data out of a PDF with python, but when I did it 20 years ago it wasn’t pretty or fun.
Yeah us we’ll do it. Our autism loves organizing shit
This is one of the things LLMs are actually pretty good at. Just don’t blindly trust its output.
If I were to use an LLM, it would not be to actually upload the PDF and generate the excel document, you’re guaranteed to have made up data if you ask it to do this. What I would do is ask an LLM to write a python script which uses OCR or some other programmatic way to extract the data from the PDF and put it into a CSV to be imported by Excel.
If the PDF has some sort of data aggregation, like a column for a sum of the data in a row, then do not include that in the CSV output, and have excel do the calculation based on the data the script imported. Then you just have to manually check that the values of that column match the PDF to know if there is any wrong data. Obviously if multiple fields are adjusted by bad OCR but negate each other, the sum column would look accurate while the bad data persists, so some more spot checking or additional aggregation would be needed to ensure confidence with the numbers.
The answer is yes, this is exactly what sites like fiverr are for.
That is, if you value your time more than your money for this, because there’s probably a way to still semi-automate it and avoid some of the work. But yes, fiverr.
You could get someone on Fiverr to do it
What does the PDF look like? Can you not OCR it into something like a CSV, and import it in Excel?
Okay finally managed to grab a table or two - sorry for the delay. But yeah this is what they look like and I just need to consolidate this data and format it better haha
Try using an LLM.
I tried to timebox myself with a Python OCR solution, but ended up spending way more time than I should’ve, without ever arriving at a good solution.
I then tried using Qwen 235B via Brave’s Leo AI, and got some okay results (I think you’ll have better luck with commercial AI models; I just don’t have an account with them atm)
Word can convert PDFs, then its “easy” to copy the data to excel and manipulate it there
The data may already be selectable text, but usually the formatting after copy paste is horrific.
The PDF is a bunch of tables mostly, but it’s tables within tables. I honestly don’t think I care to try OCR into a CSV into python or anything.
I honestly loathe these documents and reading them so I’m happy to just pay for the help
This is something that CoPilot or ChatGPT is actually good at. Depending how messy your data is you might need to several iterations to get what you want and you will need to cross check the results. If you can, use a pdf editor to delete all the information you don’t want. I use foxit for this. Foxit can also save a pdf as excel but if the tables are an image or are complex it will not format very well at all.
I can help. Feel free to send me a DM
I used to do a lot of parse-compile-report scripting, and I’m very curious about how the data is formatted in the pdf. Would you be willing to share it?
Other than that, fiverr or an LLM is perhaps your best bet.
I threw some tables in a comment a few above yours if you’d like to take a peek
deleted by creator
