topic Re: How to use Text Import Node in SAS Data Science

How to use Text Import Node

LipizinPinto — Sun, 25 Nov 2018 22:13:25 GMT

Hello, I need help with a question, please. Here is the scenario:

I will be performing a Text Mining on content extracted from Facebook, and the textual data was imported to an Excel file. The thing is: This single Excel file has a column named "Posts" and 5.000 rows filled with posts and comments. Can I import to "Text Import" node the file just the way it is, and I'm ready to move on to Text Parsing?! Or should I have one separated Excel file for EACH one of the 5.000 texts?! If so, how can I automate this generation of 5.000 Excel files containing a single text on each one?!

Thanks in advance.

Re: How to use Text Import Node

Shmuel — Mon, 26 Nov 2018 07:13:31 GMT

Please post a sample (few lines) of your input
and clarify what do you want to do with it - what output expected.

Re: How to use Text Import Node

LipizinPinto — Mon, 26 Nov 2018 21:11:30 GMT

Hello, really thanks for replying back to me.

Here is the scenario: I have a folder containing only one Excel file (below) full of tweets extracted from Twitter, and I want to use it as input for my Text Mining analysis.

Opening this Excel file, you can check a sample of its data (below), which has a lots of rows containing the textual data to be analyzed (where each row contains a different tweet, and therefore we can consider as having N documents inside a single Excel file).

My question is: When using "Text Import" node, can I inport this single Excel file and expect SAS Text Miner to understand that each row is a different text (document) to be analyzed, or am I supposed to have each one of these documents (rows) saved separately?!

For example: let's suppose inside this Excel file I have 150 rows containing differents tweets, can I inport this single file with its 150 rows or should I have 150 excel files (1 for each row)?

Hope I could express myself better this time!

Thanks in advance!

Re: How to use Text Import Node

avp — Tue, 27 Nov 2018 07:09:22 GMT

Nice elaboration of your requirement. It helped.

I think you can try with "File Import Node " from the "Sample" tab

File Import Node -> Text Parsing Node -> Text Filter Node ....etc.

In File Import Node do "Right click.." -> "Edit Variables...." and change the Role of "Tweets" variable to "Text"

That should run your Text Mining flow ,

taking each row from your excel as one individual document. (First row "Tweets" will be the variable name )

Thanks.