And what words should a perfect title contain?

I wanted to find the best word sequence to form a sucessful title in respect to a dewey class. I therefore first searched for the two most popular first words in a title. As popularity, I defined how many items were checked out with a certain word as the first word in the title. I then looked for the two most popular words that follow these first words, and then for the two most popular words that follow each of those words. I iterated that process for several iterations. It appeared to be easier to (automatically) re-execute the a query with changing parameters than writing a potentially complicated SQL query that would get all the values at once. The formatable query to retrieve the first two words:

Code:

SELECT
  COUNT(id) AS checkouts,
  LEFT(deweyClass, 2) AS deweySub,
  {wordNum},
  SUBSTRING_INDEX(SUBSTRING_INDEX(title, ' ', {wordNum}),' ', -1) as word
FROM
  spl_2016.inraw
WHERE
  LEFT(deweyClass, 2) = '{dewey}'
GROUP BY
  word
ORDER BY
  checkouts DESC LIMIT
  5


And the formatable query to retrieve each subsequent words:

Code:

SELECT     
    LEFT(deweyClass, 1) as dewey,
    LOWER('{word}') as parent,
    COUNT(id) as checkouts,   
    SUBSTRING_INDEX(SUBSTRING_INDEX(LOWER(title), '{word} ', -1),' ', 1) as nextWord
FROM
   spl_2016.inraw
WHERE
    LEFT(deweyClass, 1) = "{dewey}"
AND
    cout > "{startDate}"
AND
   title LIKE '%{word} %'
GROUP BY
   nextWord
ORDER BY
   checkouts DESC
LIMIT
   2;


I represented each word as a node in a 3D space. Each node is repulsed by any other node. The repulsing force hereby is proportionally to the inverse of the square of the distance. At the same time, nodes are connected through arcs to their predecessor and successor word-nodes. The arcs impose attracting forces on the nodes they are connecting. The nodes visualize their directionality through a small dot traveling alongside the arc from the predecessor to the sucessor word-node. Since the nodes experience friction, they eventually will end up in a stable position. Please put your final images or screen shots here with some explanations.

The visualization and the mechanics work and are somewhat pleasing. However, it is not trivial to balance out the forces in order to have the connected nodes appear with an acceptable spacing and centered in the scope of the camera.

Built with Processing 3.0.1
Source Code + Data