aggregate CSV fields explanation

T3Please_draw_a_box_around_each_individual_animal_in_rectangle_summary.csv

 subject_id : Panoptes subject ID
 median(left) : Median number of distinct cluster of markings of left-facing animals
 median(right) : Median number of distinct clusters of markings of right-facing animals
 median(away) : Median number of distinct clusters of markings of away-facing animals
 median(towards) : Median number of distinct clusters of markings of toward-facing animals
 mean_probability : The mean average percentage of people who had markings in each cluster
 median_probability : The median average percentage of people who had markings in each cluster
 mean_tool : the mean average percentage of people who agreed with the most popular direction for that cluster
 median_tool : the median average percentage of people who agreed with the most popular direction for that cluster
 
T3Please_draw_a_box_around_each_individual_animal_in_rectangle.csv

 subject_id : Panoptes subject ID
 cluster_index : zero-indexed counter for each cluster(box) identified by the crowd (each box has its own row)
 x1,y1 - x2,y2 : coordinates
 p(most_likely_tool) : the percentage of people who used the most commonly selected tool for this cluster
 p(true_positive) : the percentage of people who saw the subject and have a marking in that particular cluster. Measuring how likely it is that that cluster actually represents "something". A false positive would be something like a tree or a rock. Anything below a certain threshold gets discarded as false positive.
 num_users : The number of users who saw this subject
 
T3What_is_the_animals_pose.csv

 subject_id : Panoptes subject ID
 cluster_index : zero-indexed counter for each cluster(box) identified by the crowd (each box has its own row) 
 p(moving) : The percentage of people who thought this specific animal was moving
 p(standing) : The percentage of people who thought this specific animal was standing
 p(lying) : The percentage of people who thought this specific animal was lying
 p(sitting) : The percentage of people who thought this specific animal was sitting
 num_users : The number of users who saw this subject

T3What_is_the_animals_pose_summary.csv

 subject_id : Panoptes subject ID
 cluster_index : zero-indexed counter for each cluster(box) identified by the crowd (each box has its own row) 
 most_likely : The most likely (most popular) direction faced for that animal
 p(most_likely) : The percentage of people who held that view
 shannon_entropy : a measure of agreement amongst users for this question - e.g. if entropy is 0, everyone was in agreement. See https://en.wikipedia.org/wiki/Entropy_(information_theory)
 mean_agreement : the mean average percentage of people who voted for that option
 median_agreement : the median average percentage of people who voted for that option
 num_users : The number of users who saw this subject
 
T5Is_it_easy_moderately_difficult_or_hard_to_identi.csv

 subject_id : Panoptes subject ID
 cluster_index : zero-indexed counter for each cluster(box) identified by the crowd (each box has its own row) 
 p(easy) : The percentage of people who thought this image is easy
 p(moderate) : The percentage of people who thought this image is moderate
 p(hard) : The percentage of people who thought this image is hard
 num_users : The number of users who saw this 

T5Is_it_easy_moderately_difficult_or_hard_to_identi_summary.csv

 subject_id : Panoptes subject ID
 most_likely : The most likely difficulty for that image
 p(most_likely) : The percentage of people who held that view
 shannon_entropy : a measure of agreement amongst users for this question - e.g. if entropy is 0, everyone was in agreement. See https://en.wikipedia.org/wiki/Entropy_(information_theory)
 mean_agreement : the mean average percentage of people who voted for that difficulty.
 median_agreement : the median average percentage of people who voted for that difficulty.
 num_users : The number of users who saw this subject

T3How_much_of_the_animal_is_visible.csv

 subject_id : Panoptes subject ID
 cluster_index : zero-indexed counter for each cluster(box) identified by the crowd (each box has its own row) 
 p(complete) : The percentage of people who thought this specific animal was completely visible
 p(mostly visible) : The percentage of people who thought this specific animal was mostly visible
 p(half obscured) : The percentage of people who thought this specific animal was half obscured
 p(mostly obscured) : The percentage of people who thought this specific animal was mostly obscured or off camera
 num_users : The number of users who saw this subject

T3How_much_of_the_animal_is_visible_summary.csv

 subject_id : Panoptes subject ID
 cluster_index : zero-indexed counter for each cluster(box) identified by the crowd (each box has its own row) 
 most_likely : The most likely (most popular) opinion on how much of the animal is visible
 p(most_likely) : The percentage of people who held that view
 shannon_entropy : a measure of agreement amongst users for this question - e.g. if entropy is 0, everyone was in agreement. See https://en.wikipedia.org/wiki/Entropy_(information_theory)
 mean_agreement : the mean average percentage of people who voted for that option
 median_agreement : the median average percentage of people who voted for that option
 num_users : The number of users who saw this subject