Sentiment Analysis

Loloy

Please, I need your help. I collected tweets and put them all in one excel sheet to make a sentiment analysis. Each time I try that, it shows me this message "Process failed. API Engine error: no content to analyze [212]. Can you pls help me? I would appreciate your guidance.

MartinLiebig

Hi,

what exactly did you do to get this error?

BR,

Martin

Loloy

Thanks so much, Martin. I collected tweets for a couple of months and stored them in an Excel sheet. And now, I want to do the sentiment analysis, but each time I try that, I get that error message. I tried to use all attributes as trials within the sentiment analysis parameters, but nothing worked. Please, I would appreciate your help. Thanks!

Loloy

Please @MartinLiebig let me know if you need any more information. Thanks!

MartinLiebig

Hi,

the error comes from the API the operator is calling, not from RM itself. I suppose you have either missing or empty lines in your text.

Cheers,

Martin

Loloy

Thanks so much @MartinLiebig .Yeah, there are missing and empty lines. It is big data. So I tried to make auto cleaning in RapidMiner, but it didnt work out. Also, I tried to do the cleaning within the excel sheet by following these steps : 

Removing Blank Rows with Find & Select

1. Click Find & Select.
2. Click to Go to Special.
3. Choose Blanks.
4. Click OK, and then all the blank rows/cells will be highlighted.
5. Choose the Delete under Cells section on the Home Tab.
6. Click Delete Sheet Rows.

But the cleaning did not work properly; it took days, and even after, the excel keeps showing that it is not responding. 

Please, is there any way to clean my dataset properly to be able to run "Sentiment analysis"? Please, I would appreciate your help.

MartinLiebig

Hi @Loloy ,

did you try to use Filter Examples and remove all missing examples? Maybe also use trim first.

BR,

Martin

Loloy

Pls @MartinLiebig Something like that ? unfortunately it did not work as well

MartinLiebig

Hi,

how did you configure your Filter Examples?

BR,

Martin

Loloy

Thanks so much @MartinLiebig . It seems that I am putting in the wrong filters. I tried a random filter, and it worked, but I got the wrong results as it shows only 18 tweets. Pls, Can you help me to know the right filter? lp me

Loloy

and pls @MartinLiebig do you mean by "Filter example configuration " by filter examples rules or something else?

MartinLiebig

I would try

text is not missing

first.

Best,

Martin

Loloy

Unfortunately, it did not work. Pls @MartinLiebig what do you think 

MartinLiebig

Hi,

did you try to use the trim operator before, as suggested earlier?

BR,

Martin

Loloy

It did not work. I am trying it again now. Pls @MartinLiebig What is the right attribute filter type ?

MartinLiebig

Hi,

please use trim on your attribute and read the documentation of trim.

Best,

Martin

Loloy

I will go through the documentation, give it another try, and keep you updated. Thanks a lot @MartinLiebig

Loloy

Hi Martin,

Unfortunately, it did not work out. The main issue with my data is that it is not organized. I spent months trying to fix it, but it did not work. Thus, I am trying to conduct sentiment analysis using wordlist that came as a result of data mining, is it a good idea ? I am trying to reach to list of positive and negative words. But unfortunately, right now I am getting this message attached. Please @m@MartinLiebig help me , I would really appreciate your help and guidance.

Thanks