Imbue Merit Calculations

“Good deliberation, however, is impossible without calculation.”

This page explores a design for how site visitors collaboratively rank input that’s been published about songs, tracks and their sequences.

THIS IS IN DISCOVERY STAGE OF DEVELOPMENT

See prior versions for other and future features.

Collaborating at Imbue.im

After selecting a link from the list of Imbue Song Versions people will:

• upload new versions
  • Imbue site will curate and upload recordings
• deliberate existing songs
  • The deliberation method uses the “ushin semantic screen”, exemplified at u4u.io

Purposes

The imbue site suggests terms and factors related to the music and art presented here to encourage collaboration and deliberation while testing ushin tools for art, music and tech collaboration.

Deliberation with simple averages

Here is a set of influences on the merit given the central point – a suggestion to upload the tagged piece of music at a speed of 120 bpm. [why not do this digitally keeping the pitch … ] Maybe if just before a live performance or recording it might be worth asking those collaborating, if it were easy to get an ushin consensus.

If it were worth deliberating more fully there would be text to support the ranks that could each further be deliberated.

[diff when facts are ignored in the calculation vs when rated 0]

Example of one person’s input

Deliberation with ranked merits

This site introduces a new feature of ushin deliberation. This the value of a message, it’s level of meaningfulness, or meaninglessness, if any.

This website will encourage site visitors to deliberate the merits of suggestions made by other site visitors who publish points, such as opinions about songs and images, or have suggested actions to take, or feature requests.,

Site visitors will be able to rank art and music and deliberate input by others, with specific details.

Each participant is called a peer, and the numerical value awarded to a peer’s input is called the merit.

The merit given by a single peer for the main point of another peer is an average of the products of that peer’s evaluation of supporting points, considering the weights that the peer assigns to the kinds of points, each labeled with an ushin shape. This is then influenced with a factor for completion considering how many of the potential 7 shapes were completed.

[show example of Bob ranking Alice’s suggestion to speed up the song, show that he has to rank the song title as correct to use Carol’s weights, or anything tagged to that song title.]

Peers give merit to main points by rating the supporting points

Merits are given to published points. Peers cannot give merit to their own points before publishing them, as the merit field is not present until points are published.

• Single (supporting) point merits (SPM) are assigned by peers from 1-5, in this iteration by clicking on 1-5 stars adjacent to each of 7 fields, each representing a shape, or kind of meaning associated with the field, each representing support for the main point in center.
• Main point merit (MPM) is a product of the single point merits (SPM) and the shape weights that the peer has given various ushin shapes. When a peer doesn’t rank shapes, then the MPM reveals a simple average of all supporting points given by a peer for the central, “main point” of a semantic screen (1 to 5 stars).
• Shape weights (SW) show user preference for individual ushin shapes which define the seven fields. In this iteration the peer does this by selecting individual shapes to usher using the semantic screen.
• Qualifiers such as importance, relevance, urgency, creativity, originality, beauty and recency are suggested by the site to make it easy for users to deliberate the music. The site also provides input fields for site visitors to add shape qualifiers on the fly. In the future these valuations may be similarly “ushered”for further deliberation.
• Peer Completion Score (PCS) ranges by default from 94% to 100% based on the number of fields completed, from 1 to 7, as there are 7 fields, one for each ushin shape.

A single merit score for a select community of peers is averaged from main point merits, which, again, derive from averaged weighted single merit scores:

• Community Point Merit (CPM) shows the average rating published by all of the peers that a site visitor custom picked to weigh in about that main point. The software will store which rank goes with which shape for clarity and further deliberation.
• Community shape weights (CSW) for each shape are published by peers to rank shapes individually as with individual shape weights, above. These weights are rolled into the Peers’ Merit Score and thus into the average for all peers included. The individual shape weights will be saved for evaluation and display.
• Peers’ Completion Percentage will likely be folded into the average peers’ score because the averaged individual peer’s final score will factor in completion by default. The reason for this is to model the encouragement of more comprehensive deliberation which may be useful in future iterations of the ushin merit methodology.

Completion Factor

• Submission requires at least one field to be filled
• Score goes up the more fields are filled

Number of Inputs Filled	Score (%)
0	locally saved
1	94%
2	95%
3	96%
4	97%
5	98%
6	99%
7	100%

Peer selection of other peers to include in merit ranking

Each peer on the site selects which other peers are included in the calculations. In this iteration the display will show all of the peers which have contributed, as curated by the site. Site visitors using the ushin tools will be able to view different results by specifying which peers are included. Future iterations may allow users to rank other peers, perhaps with drag and drop with the peer list. Initially the Imbue site will consider all peer input equally. In the future this site may explore shape weights for individuals when deliberating a musical fragment. For example, the original songwriter’s feelings might be up-ranked by users or the site. Supporting factors may also, in future software versions, be potentially detailed and ranked for further deliberation.

How a single peer changes a community merit

The site will show the change in overall merit when a single peer is or is not included in the peer list findings.

• Community merit before user inputs merit
• Peer merit added
• Community merit after user inputs merit

Single Peer Calculation Examples

Where p:people; t:topics; n:needs; e:emotions; t:thoughts; f:facts and m:merit

——– EXAMPLE MERIT CALCULATIONS for a SINGLE PEER ———>

• Average Weighted Score (AWS): Average score (1-5)given by a peer across fields 1-7,
• Shape Weight (ShW) Factor weighted by peer according to the importance assigned by the peer.
• Completion Factor (CF): Number of fields completed by the peer. For example, “7 (100%)” means that all 7 fields were completed.
• Adjusted Merit (AM): Overall user score that takes into account the weights of each score and their respective completion percentages.

Single and Community Merits

Example results. Compare the first and last columns to see how a single peer evaluating a point changes the community merit for the selected set of peers.

Community Metrics:

• Peer Count: Number of peers contributing to the community metrics.
• CSB (Community Score Before This Peer): Average community score before considering the contributions from the user, based on previous inputs from the community.
• CSA (Community Score After This Peer): Updated community score after integrating the user’s contributions.
• CSB and CSA are calculated based on community data. While formulas may vary, on this site, Imbue, final peer scores will include peer and completion weights and then average final scores by the number of peers contributing merits to the point.

Custom weights

Custom weights (one’s own and group rankings) – future feature request . In the future peers could, for example, remove a shape’s valuations from results, or re-calibrate point merits using different factors, applying different limits and including kinds of fact citations or peer qualifications.

For metrics related to scientific journal articles see dev.ushin.net/metrics

Move to dev site, as this is not applicable to Imbue:

Evaluation of Metrics for journal articles, focus on scientific publications: