Help about: Tag cloud - TF - Groupby - Term (Urgent)

Hi folks! I can’t figure out why my word cloud looks like this. When i see output table, i wanna delete some of them but it’s not allowed I guess. Only option is rule engine…
Please, Kindly asking you to help me with my project. I work so hard but still could’nt solve that isshue. Thank you already…

knime_project_first3.knwf (94.2 KB)

Hi @matals

No worries, you are almost there. First step do only a GroupBy on Term, so remove TF-rel

Next is to decide what you want to do with the Relative Frequency (TF rel). Do you want a mean, sum or just a count (depends on your use case). And use this variable as a value column in the Tag Cloud.

Then you are using the Tags to String node, I think the Term to String node is a better option here. So it will remove the ‘special’ character in the new string.

Did you check the N Chars Filter and Stop Word Filter from the KNIME Text Processing nodes, it will help you to create a more readable wordcloud.

Hope this helps

gr. Hans


Thanks a lot indeed Hans, you saved me. Maybe It’s not that hard but I’m a beginner so couldnt slept for 2 days to fix that problem. Tbh, I told my group friends to let me use Knime I can handle it, but If I can’t get the result, I have to quit and do the manual way, dont want to give up. But theres no any time left so, please can you help me about my current situation? I do what you said but my tag cloud is still not look like the way it should be. (1. photo)

my work
knime_project_first3.knwf (94.2 KB)

Hi @matals

Although I’m not familiar with your use case, I think you added the wrong column name in the Strings to Document node. I added a Stopword Filter node (Turkish) and the N-Char filter node, to filter out all words with 3 or less characters.
I adjusted the GroupBy node and JavaScript based Tagcloud node to your stream for (i.m.o) better looking WordCloud (but I’m in doubt if you can create exactly the same image as you provided in your post). See knime_project_first31.knwf (170.9 KB)

gr. Hans


I probably will use rule engine to remove some of the words if it is okay. And you teached me better then any. You are great. Thank you!

1 Like

UPDATE: For those who reads these post, I have an information to share!

For removing, I didn’t use the nodule i said. Instead of rule engine, I set a table creator right before stop word filter and entry words that I don’t want to see in my word cloud. It work better.

Here’s the current version of my work:
habermetni_analizi.knwf (108.1 KB)

Thanks a lot to @HansS again for quick, useful reply and shaping, I was about to give up when it was so close :slight_smile:


This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.