After building the qr data we thought it would be a good idea to have a play with Watsons Personality insights on node red. This is where things got interesting, we found it the service does not just require 100 words in order to scan within those words their at least has to be 70 matches to its personality algorithm in order to get a result.
After some tinkering with different text bios we found online we got some data out of it and outputted it into a debug node. We had assumed we would get some names with a percentage but what we actually got was a very long Jason file with a bunch of children branches. In fact the file was so large both the debug node nor mqtt would output the file and it was only showing us the first 4 lines which was unusable.
After this we decided to look into how the personality trait node works. Looking at the documentation the node uses a system called “The big Five” Personality module which was developed by Costa and Norman. As the name suggests this system goes of 5 main personality traits which get a score off of their sub traits. The Five traits are
Like I said each of these traits have got a further 6 traits underneath them. The trait score of the big five is made up of the smaller scores making for an overall score. For our project we are only going to look into the Big Five, but for future development we would like to go deeper into the personality tree. As this will give the project more depth and allow people to meet on a deeper level compare to just reeling on the Big Five traits.
To use this system properly we will have to see how we can view the Jason file in order to see the correct data. Once we can view the Jason file we can begin to work out how to extract the sections on the Big Five and integrate that into our project.
After some tinkering with different text bios we found online we got some data out of it and outputted it into a debug node. We had assumed we would get some names with a percentage but what we actually got was a very long Jason file with a bunch of children branches. In fact the file was so large both the debug node nor mqtt would output the file and it was only showing us the first 4 lines which was unusable.
After this we decided to look into how the personality trait node works. Looking at the documentation the node uses a system called “The big Five” Personality module which was developed by Costa and Norman. As the name suggests this system goes of 5 main personality traits which get a score off of their sub traits. The Five traits are
- Agreeableness
- Conscientiousness
- Extraversion
- Emotional Range
- Openness
Like I said each of these traits have got a further 6 traits underneath them. The trait score of the big five is made up of the smaller scores making for an overall score. For our project we are only going to look into the Big Five, but for future development we would like to go deeper into the personality tree. As this will give the project more depth and allow people to meet on a deeper level compare to just reeling on the Big Five traits.
To use this system properly we will have to see how we can view the Jason file in order to see the correct data. Once we can view the Jason file we can begin to work out how to extract the sections on the Big Five and integrate that into our project.