Data Divination: Technical Realisation

The technical element of this project intends to create a framework by which the pattern discerning capabilities of both the machine and the human can be captured within and analysed by a database architecture.The end goal is to have a rich database of the experience of reading tea leaves. A data mining process will be executed upon this database to investigate the difference between intuitive knowledge and data mining knowledge.

In order to successfully data mine any area of experience or enterprise a sufficiently deep amount of data is required. Thus the most pressing issue is finding ways to extract data from the practice of tasseomancy.

Work on this project has proceeded along the following arcs so far:

1.) – Conceptual Model Of Divination

An attempt at engineering a conceptual model of what the tea leaf reading process involves. This will distill the practice of tea leaf reading into a set of distinct fields of tables which will constitute a database. Thus far there are two tables: one which attempts to capture the data incurred while engaging in tasseomancy and a second table representing the computer vision analysis of the tea leaf reading process.

The development of the database is intended to be informed by engaging people with the practice of tea leaf reading. In that regard this database will differ markedly from the usual data mining process. Typically an enterprise will have a great deal of extant data pertaining to their activities and the challenge is how to work this data into a relational database. Here the relational database architecture will grow with tea leaf reading and those who practice it.

The ‘human vision’ table of the database will reduce the whole experience of the tea leaf reading to a set of simple questions, submitting the process to a set of sterile questions which quantify the experience in manner suitable for database normalisation.

2.) Machinic Vision

Ultimately the machinic pattern discerning element of this project will be fulfilled by data mining the sum total of the above database. It will be a while before that point is reached, and in the meantime machine vision encompasses an interesting field of it’s own (in that it pertains to machine learning and also quantifies the real world in terms of patterns more amenable to the CPU, i.e. computer vision is markedly inferior to human vision).

The technical elements thus far encompasses:

A webcam recording the tea leaf drinking from the bottom of a glass cup.
The captured webcam footage will be split into frames for pattern analysis. Pattern analysis of the footage whilst in motion is also hoped for.

The output of this research has been pleasingly enigmatic so far (see the media gallery for more)


An ongoing record of research into this project can be followed on my blog {LINK}