1-percent who stand with the 99-percent: more meta analysis

Though I may appear to be getting obsessed now, I imported the 97 stories from 1-percenters who talk about who they are and why they stand with the 99% of the world in forcing economic change into a program called SenseMaker(R) developed by Cognitive-Edge.com. This allows me to analyze these patterns in stories:

1. Story redundancy,
2. Categorize story focus,
3. Analyze clusters,
4. Look how the order that stories have been coming in to the blog over the last 2 weeks affects the themes discussed by 1-percenters.

Story Redundancy

Each point on the plot represents the position of one story. From left to right below, stories with a lot of unique words (not found in any of the other stories) can be separated. From top to bottom, stories are positioned based on how much of the text matches words related to one of 11 larger topics (ranging from basic needs to the freedom to pursue one’s dreams):

The “most unique” purple story turned out to be in French. So duh, of course it is a unique outlier. Taking the orange stories together, here is what you find in their opening words:

Compare that with the red group of stories which are the most redundant:

Eureka! It works! Notice that every one of these stories from the redundant group is about family. More specifically these are about how birthright, inheritance, trust-funds, and other mechanisms designed to secure wealth as it passes from one generation to the next. Other words that come to mind are dynasty, lineage, and legacy. So while not all of the 97 stories are about this, the largest subgroup appears to highlight the root cause behind why the rich keep getting richer: once wealth accumulates within a family at a sustainable level, it doesn’t leave.

Another fuzzier method I tried was to associate about 70 frequently used words from the gephi-word-map in my previous analysis with 11 categories:

Story Focus (as a needs hierarchy)

1. Basic needs
2. Family
3. Knowledge or Education
4. Root causes of the problem
5. Respect, Justice, or Dignity
6. Self-esteem or Opportunity
7. Creativity and Free expression

(Note: some categories did not match any stories and were rolled up into these others. In the future, one of my side hobbies is developing lexicons like this to translate any story into a hierarchy for analysis. )

Stories could be assigned to one or two of these categories and then placed on the overall chart:

Each horizontal band of color corresponds to one of the focus areas. The largest group of stories fell into the Root Causes category (29 of 97):

Or the Gephi version:

It would appear that the root causes of our problems have something to due with wealth in families. I’d interpret the rest of these phrases as what people want to do about it: We want to live in a world… where people can follow their dreams [if / and?] the richest 1% give back to the rest. This is partly speculation, but I have read all the stories, and from my memory the “worked hard” phrase was always about how storytellers or their parents /families got rich to begin with.

Story order

I also analyzed the order in which these 97 stories were submitted, but no interesting patterns have emerged yet. If I did this again in a month with 500 stories, you might find a topic drift as people tell stories partly to respond to the media narratives that misrepresent them. Stay tuned.

Cluster analysis

When we have thousands of stories, I can run a self-similarity organizing program to partition stories into groups that will have a similar topic, much like the redundant group I mentioned above. The difference is that when you repeat this analysis many times over, the definition or unique character of each cluster of stories within the whole can change over cycles and display some “strange attractor” behavior. I’m seeing whether this teaches you anything interesting using the 25,000 stories collected about communities in East Africa.

Final note on SenseMaker(R)

Normally the SenseMaker(R) software is used to analyze stories that have an associated signification-framework. This time I am only using the text of stories from a website and the order in which they were posted-  nothing else! This might surprise people who spend a lot of time using long surveys to collect data – that you can learn something without structure in the raw data. Text is often overlooked as a form of information because people think it is purely qualitative. But if there are good programs out there to organize the stories into information, you can use learn from it.

Read part one: 1-percent stories meta analyzed.