Challenge #227: /*comments*/

[AYXAcademy]

A solution to last week’s challenge can be found here. 

As you know, weekly challenges can be challenging...but did you know that they can also be challenging to create? We try to make them in a variety of difficulties, using different tool palettes while accomplishing different tasks. Some of our challenges focus on common objectives from the community and others look for unique tasks to try using tools in new ways. One of our greatest assets when trying to come up with new challenges is feedback from old challenges. See where this is heading?

In this week's challenge, we need to investigate the posts from our previous weekly challenges. The WCData.yxdb file contains comments from challenges 100-223. The first part of this challenge is to format the data into something legible. (don't worry about matching the number of lines exactly; just try to get the text into a similar form)

As an optional second part, use the new Text Mining tools to find the 10 challenges with the highest percentage of positive and negative comments. (Requires 2020.2 and Intelligence Suite. The Intelligence Suite is a new add-on to Alteryx Designer that includes code-free Text Mining and Machine Learning tools to help you solve questions using advanced analytics. You can download a trial of the Intelligence Suite if you'd like to try it out for yourself. If you've previously downloaded a trial of Designer, request access to the tools here.)

[patrick_digan]

Spoiler

I used Regex, and I also eliminated the original post.

[paulfound]

Sorry to be that guy but... you have 16,201 records the output file has 16,319 (I got the same when I did it) 

EDIT:

I'll shut up, I just got 16201 as the answer my bad! 😞

[Blake]

Not sure if I nailed the exact results as the provided answer but I feel good enough about it to do text analysis on. 

Spoiler

I didn't include the original posts or any without comments. 

[patrick_digan]

@paulfound They were including most (but not all) of the original posts in their solution, and I just decided to get rid of the original posts all together. 

[paulfound]

Spoiler

 I got 15,084 after taking out blank messages.

I would love to do step 2, but don't have the insights pack yet.

[Kenda]

Spoiler

[etzele]

Not 100% the result shown in the solution but I guess one can make it count

Spoiler

[johnemery]

This was an interesting challenge...

I believe my solution for cleaning the data is sound, but I cannot match the number of records in the solution. This is another example where we really need more concrete instructions, otherwise we are left to guess, and I don't believe that is in the spirit of these challenges. I spent over an hour going down multiple avenues to get the correct number of records without success. In the solution I've provided here, I have 16,326 records while the solution has 16,319; I cannot get it any closer, and I'm not sure what I'm doing wrong.

Spoiler

[atcodedog05]

Here is my solution. This challenge is more of a free style approach.

Spoiler