joe-the-ark
commented
1 year ago Let's do a test with the game's data for ubung-2. Let's start with the game: NISIHOFTAKEOFF. Game is a 11 player game (if I remember correctly). Ubung-2 stores ONE value per player. The answer to the question <How good is your perceived team performance?> On a scale from 1-99.
The average value in this game for Ubung-2 across all players is 53 – with 10 being the lowest value and 79 being the highest value: If we are able to retrieve this data, for that game; then compare this value with the result for ubung-4 (the psychological safety value: 69 in this case) > I am interested to see, how far these two values are apart. To analyse on a game by game basis, how far or close the values for ubung-2 and ubung-4 are across all games.
One Hypothesis could be: these two values (between ubung-2 & ubung-4) are the further apart, the further the values for ubung-2 are away from one another (here the spread between high and low values for ubung-2 between 10 and 79 is kind of extreme). Or: these values are getting increasingly lower (ubung-2 average 53 is kind of low), the further the values for ubung-2 are apart.
P.s.: Maybe one more thing - can we filter out the results from all the games that contain TEST in their name?
zdavatz
The focus of this update is in analysing & visualising new data - less so on working with 'old' data. I changed the ark significantly in the last weeks. The basis for my research shall be all the data from here on.
For the ark data, I expect something simple, to start and learn data analysis myself. It has been such a fruitfull process to learn the basic programming techniques thanks to @zdavatz, @linker03, @reindexer and @oak-innovation, I can now develop the ark myself - how great is that? I would like to embark on a similar process to learn the ropes of data analysis and data visualisation. What do you think would be a good start? I think: Export the ark backend-data with a time-stamp, probably into excel? Some simple sorting functions (e.g. display only games after 1.8.2023) and some simple lists to start (what are the key terms that have been choosen the most), display of some key indicators per game in a list (e.g. psychological safety value, tension value, performance value), check if a game was finished and only include data of finished games.
To start I am aiming at analysing across games. We have several key measures: 1) psychological safety score, 2) performance score, 3) tension score for that team. I want to display them per game to being able to compare them across games. One assumption that I observed: psychological safety score (e.g. 67 out of 99) & performance (e.g. 69 out of 99) are usually identical. While the tension level usually is in the area of 100 - psychological safety score (100 - 67 = 33 -> tension level). Just an example.
In I later stage, I maybe could make more use of an analysis per game or even per person in a game to replace the current pdf that is generated at the end of the game per user. Currently this would be nice to have, but has no high priority.
I just realise that (some of) these values are computed in the ark software. I am just wondering, can we re-use this code, e.g. calculating the tension score for a team in the headmap, or do we need to re-write?