BibExcel is a bibliometric tool-box developed by professor Olle Persson. This manual is based on a manual written by professor Olle Persson. Download this manual as a pdf-file from here.
Preparing the data
Data that is collected from i.e. Web of Science, WinSpirs/Silverplatter or Endnote must be converted before it can be analyzed. Picture 1 below shows a bibliographic description in such a format that BibExcel can read it. Every field must end with a ”spike” and the last field must end in a ”double spike”. Field tags are given with two letters, followed by a – and a space.
FN- DIALOG(R)File 7:Social SciSearch(R)| CZ- (c) 1999 Inst. for Sci Info. All rts. reserv.|
TI- Collaboration and author productivity: A paper with a new variable inLotka’s law|
AV- ABSTRACT AVAILABLE|
AU- Gupta BM (REPRINT); Karisiddippa CR|
CS- NATL INST SCI TECHNOL & DEV STUDIES,SCIENTOMETR & INFORMETR GRP/NEW DELHI 110012//INDIA/ (REPRINT); KARNATAK UNIV,DEPT LIB & INFORMAT SCI/DHARWAD 580003/KARNATAKA/INDIA/|
JN- SCIENTOMETRICS, 1999, V44, N1, P129-134
PU- ELSEVIER SCIENCE BV, PO BOX 211, 1000 AE AMSTERDAM, NETHERLANDS|
SF- CC SOCS–Current Contents, Social & Behavioral Sciences|
SC- INFORMATION SCIENCE & LIBRARY SCIENCE|
AB- The paper explores the possibility of using a new variable represented by the number of collaborators per author as a substitute for the number of papers in Lotka’s distribution to predict the productivity strata.On the basis of a case paper in theoretical population genetics it is concluded|
CR- FELSENSTEIN J, 1981, BIBLIO THEORETICAL P
LINDSEY D, 1980, V21, P111, SOC STUD SCI
NICHOLLS PT, 1988, V39, P287, J AM SOC INFORM SCI
NICHOLLS PT, 1989, V40, P379, J AM SOC INFORM SCI
PRICE DJD, 1966, V21, P1011, AM PSYCHOL
QIN J, 1995, P445, 5 INT C INT SOC SCIE
ROUSSEAU R, 1988, V39, P287, J AM SOC INFORM SCI
ROUSSEAU R, 1993, V49, P409, J DOC||
When you save a file from WoS it is automaticly named as Savedrex in txt type. The file is in other words saved as a text file.You then need to re-save the file. Open it in Microsoft Word, or some other similar program. Save it as a plain text file. In the next small window that opens, mark the box “Insert line breaks” and make sure CR/LF is chosen. This is a good time to change the name, so that the original file will not get messed up even by an accident.
Next you need to open this new file in BibExcel and convert it to Dialog format.
1. Mark directory, in this case drive C:
2. In Select file here, open the folder the file is located in and choose the file from the right.
3. Go to Miscellaneous (Misc) and select Convert to Dialog-format and mark Convert from Web of Science. Click OK and the file is converted to Dialog format 4 and saved as a .doc file. The original file is left untouched.
You can view these two files by marking them from the space above ”View file” button, and clicking on ”View file”. You can see the file in “The list”. You can compare the before and after converting files, and see how the delimiters, spikes and dubble spikes are inserted.
1. Open your data file in BibExcel (.txt). Choose “View file” to open it in The List.
2. From ”Frequency distribution”-box, choose from the drop-down menu ”Whole string” check the checkbox labeled ”Make new out-file” and write in Old tag-field DE (Descriptors or what ever label you use). This will create a new file, a .oux -file.
3. Choose the .oux -file and open it to The List by clicking on “View file”. Then, from the “Select field to be analysed…” -box, choose from the drop-down menu ”Any; separated field” press on the Prep-button. This will create a new .out -file, where all the keywords are listed by cases.
4. Open the .out -file in to The List. Choose from the Analyze-menu, ”Frequencies, using outfile-type”. This will create a .frg -file where the keywords are listed with their frequencies (how many times the keyword has appeared in all the cases).
5. Copy these by clicking on Copy and move them to Excel. In Excel sort the keywords descending by their frequency. Then copy the most frequent keywords and their frequencies back to BibExcel.
If you choose too many keywords, the maps will be too big to draw or to interpret the results properly. If you choose too few keywords, you probably will not get any interesting results. So choose wisely. Also make sure to choose some meaningful frequencies. If for example the ten most frequent keywords all have a frequency over 50 and the rest of the keywords have only appeared a couple of times each, then it is probably meaningful to take only the ten most frequent keywords to further analysis. Around 40-50 keywords makes still quite nice maps. A maximum is perhaps around 80-90 keywords.
Clear The List by clicking on Clear. Paste keywords into BibExcel and to The List by clicking on Paste button.
6. Now mark the .out -file from the file list and choose Analyze –> Co-occurrence –> Make pairs via listbox. Answer NO to the first question and OK to the second. This will create a .coc -file.
7. Continue with the .coc -file and choose Analyze –> List units in pairs.
This will result in a .ccc -file.
8. Open the .ccc -file into The List and mark the .coc -file from the filelist.
Choose Analyze –> ”Make a matrix” and answer the question in a way that suits your research. Probably, OK to the first question, YES to Lower left matrix, YES to sort the columns, NO to sort them numerically and finally OK.
9. Then choose Analyze –> Make a map/Systat… and answer OK to all the questions.
10. In the opening window write ”Submit map”. Then back in BibExcel choose Analyze –> Show map.
11. Choose colors as you please and return to BibExcel by double-clicking on the map.
12. Open the .lab –file into The List by clicking on View File. Then click on ”Show map”.
13. In the map click on Labels.
14. Return to BibExcel by double-clicking on the map. Clear The List and Paste the frequencies that you copied earlier from Excel.
15. Return to the map, click Show map, check the checkbox Zoom+ and click on Circle size. This will change the circle sizes to match their frequencies.
And you’re done. Congratulations!