PDF Import – What is the Business ValuePDF Import – Business Value Discussion
What:
Import data from PDF using new PDF Import (use the PDF Import activity to try this feature).
Suppose you are interested in data that you have found embedded in a pdf file and you wish to further explore the data.
Type of companies to target:
This kind of data can come up in any company but cannot use in this way easily. Think about
- reports generated with other software
- Word documents with tables embedded, then saved to PDF because of validation requirements (not able to change something on the document)
- Reports from lab data not entered in spreadsheets but a lab program saving directly into PDF documents
- Log Files saved in PDF Files
- Machine data written into PDF files (e.g. warning summaries, log summaries, operational change logs, …)
- Official public reports (or reports available for download as PDF anywhere in the internet) you want to use the listed data for your own purpose
- Public lists of spec limits or properties for any kind of product, demographic information, …
- This could be pharma (lab data), semiconductor (machine data), manufacturing (log/machine data), automotive (Quality standard spec limit listings for OEMs, public reports for car reliability/run time data, …)
Buying Influencers to target:
- Usually the UBI, as there is the pain! (I saw customers typing in all the PDF data from their lab reports! Cumbersome and time consuming!).
- This can be a typical Excel user
- Or a JMP user having to use data in PDFs and therefore using Excel to get the data
- The manager of the UBI, too, will see the benefit if more can be done within the same time removing such burden.
- Most of the time no one will do anything with these PDFs other than reviewing and making conclusions without proofing the data and listed statistical conclusions
àNew field of gaining information, insights! (All about being a detective
Questions to ask (besides “Do you have data in PDFs?”):
- How is the length and complexity of the PDF data you want to convert to e.g. Excel?
If you constantly convert lengthy or complex documents from PDF to Excel, the most seen online recommendation is “you may want to consider using a PDF converter such as Able2Extract”
- How often (or how many times) do you need to convert PDF files to Excel?
If you find yourself constantly carrying out the process of converting a PDF file to Excel, the most seen online recommendation is “a PDF converter (such as Able2Extract) may come in handy”
- Both questions can now be answered differently: Use JMP and you have in addition an interactive, dynamic analytical package helping you to become more efficient and better in your path to get from data to information and to conclusions. And you can automate this process if the PDFs look the same over and over again using the generated script.
Typical argument “I do this with some standard Excel plots and a Macro”:
- This is simply not possible. You have to type in the data from the PDF document by hand:
Step 1: Open the relevant PDF file and select and copy table. Step 2: Copy this into Word and convert to a table. Step 3: Paste this into Excel. Step 4: Manipulating the Excel table
- Other option would be convert PDF File To Excel Using Microsoft Word Online (again several steps)
From a blog Excel off the grid in 2016: https://exceloffthegrid.com/copy-a-data-table-from-pdf-into-excel/: A perfect solution?
“I have even used Adobe’s own PDF to Excel converter, and even that requires a lot of manual adjustments. Therefore, I don’t believe that a perfect solution exists for this problem. But, hopefully one of these workarounds has provided a reasonable solution and has saved you a lot of re-keying time.”
Other alternatives/tools:
- There are of course other tools available, but most of the time you must pay for it (quite much), like these ones:
Personal Win/Benefits:
- Reduce time spend on typing/importing data in other ways (may also save cost for PDF converter)
- Be able to use information in these PDF files to inform other analysis you were not able to use before
- Do more with the same man power in the same time -> become more innovative
Examples:
In JMP 15 EA use File->Open to bring up pdf in the pdf file wizard. Or additional example files added in the zip file “SE_PDFImport_Discussion_MD2019.zip”. PDF_
Caution/Some improvement suggestions: (July 2019)
- Empty Values:
- Date Format:
- It is as in JMP still an issue that the default import date format JMP uses is m.d.y! Why not have a field to set the input format in every import platform??? covered partially in http://sww.sas.com/cgi-bin/quick_browse?defectid=S1475858
- The issue occurs if the first entry in the column is not clearly differentiable like month and day is lower/equal than 12. In Europe most of the time the default format is d.m.y and you will always have empty values when importing due to JMPs default setting you cannot change! Partially covered in http://sww.sas.com/cgi-bin/quick_browse?defectid=S1516886
- Concatenation:
- “Concatenate all tables into one” vs. “Concatenate tables with matching column names” :
- If you concatenate into one table PDF import works differently than JMP. If the column names will differ, JMP will add a new column and fill the first ones with empty cells. In PDF import it will concatenate whatever comes which is different to what JMP does and seems to be more correct.
I understand that this will force the first table to act as header table, then all the others will be used simply as data. But this is not clear to me in first sense from reading concatenate into one table, I would expect the concatenation behavior of JMP.
- I understand that the JMP behavior should be simulated by the “Concatenate tables with matching column names”, which is really nice, however if I want to concatenate tables with some differing column names (but not all), I would need to do this in JMP after importing the tables. It would be nice if this could be done already in the Import wizard
- Concatenation all tables into one, but with multiple header rows:
- If a PDF table has a header with multiple rows, JMP’s PDF import will detect this and make a separate table out of this header. The issue is, that the last row which is part of the header cannot be put into the header using “number of rows to use for header” as one row must act as data it seems. Even if you choose to select Minimum row count to 0.
- To work around you have to manually select the range of the full table and then you will be able to have a the initial first table as a complete header.
The document summarizes my thoughts on PDF Import feature, including
- Types of companies to target
- Buying Influencers to target
- Questions to ask
- Typical arguments "I do this with Excel plots and a macro"
- Other alternatives/tools
- Personal wins/benefits
- Cautions/improvements on the way incl. defects
