many-speech-analyses / many_analyses

Many Speech Analyses
https://many-speech-analyses.github.io/
Other
1 stars 3 forks source link

flow chart #284

Closed timo-b-roettger closed 1 year ago

timo-b-roettger commented 1 year ago

comments on flow chart. Regarding color scheme: I used the orange from fig 3 and used complementary colors

image

timo-b-roettger commented 1 year ago

5-color contrasts with qualitative colors contrasts is very difficult

timo-b-roettger commented 1 year ago

alternative color scheme (Just realized some of the color mappings are off) image

Also looking at this there are things that do not add up: how can duration be operationalized as a point or unclear?

jvcasillas commented 1 year ago

I think I definitely prefer the suns themed colors over the xmas colors. With regard to your question, I'm not sure about "unclear" but I assume "point" just means they measured the duration and submitted a single data point per item or whatever. Looking at the plot, "point", "relation" and "unclear" seem to be the only possibilities, right? How else would you do it?

jvcasillas commented 1 year ago

Also, in the second plot, the colors coming out of the options on the left dont match with the actual colors in the "what" block. For example, the "intensity" block is red, but the colors coming out are green. EDIT: I just noticed that you already caught this. Ignore me.

stefanocoretta commented 1 year ago

I prefer the first plot colour scheme. The second one is not colour blind friendly.

Screenshot 2023-01-06 at 15 42 49 Screenshot 2023-01-06 at 15 42 41
timo-b-roettger commented 1 year ago

I think I definitely prefer the suns themed colors over the xmas colors. With regard to your question, I'm not sure about "unclear" but I assume "point" just means they measured the duration and submitted a single data point per item or whatever. Looking at the plot, "point", "relation" and "unclear" seem to be the only possibilities, right? How else would you do it?

@stefanocoretta @jvcasillas this is probably really inconsistent: I have never filled in anything for durational measures which is probably accounting for the many paths going into "unclear". "point" i interpreted as a singular time point which did not strike me as fitting for duration.

I would like to fix this, but it might affect the numbers here and there. I would like to either (a) code duration as "interval" (unless it is truly a ratio), (b) or all as point as Joseph suggested. Do you anticipate any major consequences or would the numbers neatly be pulled from the sheet?

jvcasillas commented 1 year ago

I guess (a) makes the most sense in hindsight. The major consequences... all of the reported values (in prose and in tables) would be different, but that would be taken care of when we reknit the document, right? @stefanocoretta can you think of anything else downstream that would have to be fixed?

timo-b-roettger commented 1 year ago

I have now changed the coding sheet. So we need to run the whole thing again and double check sensitive areas. I have changed the operationalization of duration to "interval". I also changed "ratio" to "relation" for other categories because ratio and relation was used synonymously. @stefanocoretta objections or unforeseen issues with that?