Tinderbox Training Video 56 - Sentiment Analysis & Analyzing Tweets with Action Code

Tinderbox Training Video - Sentiment Analysis & Analzying Tweets with Action Code

Level Intermediate
Published Date 7/24/21
Revision 1
Type How To
Tags 4CKMEl, 4Cs of Knowledge Management and Exchange, Action Code, Privacy, Self-sovereign Identity, Sentiment Anaysis, Tim Berners-Lee, Tinderbox, Twitter, aTbRef’
Video Length 13:53
Video URL Tinderbox Training - Sentiment Analysis & Analyzing Tweets with Action Code - YouTube
Example File TBXL - L Sentiment Analysis & Analyzing Tweets with Action Code.tbx (1.9 MB)
TBX Vesion 9.0l Sentiment analysis is available in seven major languages using macOS 10.15 and later.
Instructor Michael Becker

In this lesson, we take a look at sentiment analysis in Tinderbox (requires Tinderbox 9). Tinderbox has two system attributes, $Sentiment, and $Sentiments, which are used to analyze the sentiment of every note. That is, we can now know if the note’s sentiment is negative (-1), positive (+1) , neutral (0), or someone in-between. Tinderbox does this by leveraging the natural language processing capability of the macOS.

Tinderbox analyzes a note’s sentiment by analyzing both the $Name and note $Text. I show you how, with action code, to remove the $Name from the analysis so that you can also look at just the sentiment of the $Text.

In addition to showing you how to use the sentiment feature, I also show you in this lesson a practical example of sentiment analysis. I show you how to export the tweets for someone from Twitter to Excel (using https://www.vicinitas.io/), import the tweets into Tinderbox (via a CSV file), and use action code to transform the tweets and calculate the sentiment for each tweet and for all the tweets. In this case, I am looking at the tweets of Sr. Tim Berners-Lee (https://twitter.com/timberners_lee), the creator of the web (NOTE correction for the voice-over, I mistakenly say Internet).

TBXL - L Sentiment Analysis & Analzying Tweets with Action Code

Table of Contents:

00:09 - Introducing Sentiment Analysis

00:36 - Adding a note with next

01:05 - Displaying the sentiment attributes

02:35 - Look at the sentiment without the title

05:58 - Analyzing thousands of tweets

07:57 - Importing CSV

08:25 - Changing note names with action code

09:14 - Add prototype

10:07 - Apply prototype/update inheritance

10:56 - Analyzing all the notes at once

13:49 - Call to like & subscribe


Hi All, the link to the sample file is not fixed. I’d not properly escaped the pipe in the markdown table.

Also, had a good comment on the YouTube channel. To have Tinderbox update a note sentiment and sentiments calculation you need to click away from the note and then click back to it to give it focus. Once the note has focus Tinderbox will re-calculate the sentiment and sentiments.

Nice Job Michael. I especially liked the outline and timecode summary segment. Appreciate your work.


1 Like

Doesn’t work for me. Restarted TB, restarted Mac, restarted myself. :slight_smile:
Nothing worked out.


Hi, Andrei!

  1. Sentiment analysis requires Big Sur.
  2. Sentiment analysis is available in right now in seven major languages. I think Russian is among them.
  3. In Tinderbox 9.0.0, Sentiment analysis operates in the system’s dominant language, which might not be the dominant language of a particular note.

We’re working on near-term improvements to #2 and #3.

Hey Mark, I took a look a the vidoe, might there be a hardware requirement too? He is on a

13" 2015 machine.

Ups. I didn’t blur my serial number. :scream:

Post updated. You might want to delete your YouTube video.

I just upload shorter version of that video https://youtu.be/UnAq5ScqdOg
I tried with and without Russian actually. None of combinations worked.
It is not vital feature for me. I’m just wondering what is causing problem.

1 Like

What are the your language settings in system preferences?

And in which language are your notes?

Language & Region
Preferred languages:

English - Primary


I tried to make a completely new note/document in English. BTW in Russian notes Sentiment doesn’t work too.

In this case, sentiment analysis should operate in Tinderbox 9.0.0 in English.

Idea: post a small document here, so we can see what’s happening.

Here it is.
Тестовый файл.tbx (103.9 KB)
Test.tbx (74.2 KB)

Here are the sentiments I see in Test.tbx on my (English-speaking) machine:

Hm-m. Mystery… So we can conclude that this function will not work on some Macs even with Big Sure (v11.5).

We’ll continue to investigate. It might be a system problem, or a shortage of disk space, or something else…

Maybe this will help to improve the system.
For English, the choice of the region was critical. I had Russia.
When I switched the region to the U.S., the sentiment analysis started working. My default language has always been English.

No switching helped for the Russian language. Sentiment analysis doesn’t work with Russian for now. I tried changing the default languages, regions. Nothing works.

Thanks! We’ll check with Apple whether this should work now, or whether it’s coming.

Dear Mark,

That may explain why I cannot get sentiment analysis on TB9 to work on my Mac (11.6) that is set on Swedish even if I run all programmes in English. If that is the case, do you know of any workaround?

Best wishes,

What is the preferred language on your system (System Preferences: Language And Region)?