ConservationInternational / ShinyCam

Basic Analytics for Camera Trap Data
6 stars 1 forks source link

Administrative statistics #31

Open gchoy opened 7 years ago

gchoy commented 7 years ago

Problem: Calculate and display metrics relating to how well CI staff is identifying animals, because animals in photos are currently manually identified by members of CI staff (as opposed to via an automated computer vision algorithm).

Desired Solution: Count catalogers, image uncertainty by cataloger, Count of images per cataloger ** results should be both cumulative as well as filterable by user defined date range. Metrics:

The metrics should be displayed in a new tab labeled Operational Statistics. New metrics should be displayed in tabular form. For example:

Cataloger # of images Project ID
Name number
Cataloguer # of Images Level of Uncertainty
Person 1 number absolutely sure
Person 1 number not sure
Person 1 number pretty sure
Person 2 number absolutely sure
Person 2 number not sure
Person 2 number pretty sure

Columns names from csv files:

Deployment CSV Cameras CSV Images CSV
Deployment ID Project ID Project ID
Event Camera ID Deployment ID
Array Name Make Image ID
Deployment Location ID Model Location
Longitude Resolution Serial Number Photo Type
Latitude Resolution Year Purchased Photo Type Identified by
Camera Deployment Begin Date Genus Species
Camera Deployment End Date Uncertainty
Bait Type IUCN Identification Number
Bait Description Date_Time Captured
Feature Type Age
Feature Type methodology Sex
Camera ID Individual ID
Quiet Period Setting Count
Restriction on access Animal recognizable (Y/N)
Camera Failure Details individual Animal notes
gchoy commented 7 years ago

Administrative Statistics script with desired metrics written by Sarang Potdar and Neslihan Tuzun. Link to pull request: https://github.com/ConservationInternational/ShinyCam/pull/38

gchoy commented 7 years ago

Count of images per cataloger output:

Project.ID Photo.Type.Identified.by n
1 ChedaJewel 41564
2 ChedaJewel Andre C 1
3 ChedaJewel Andre Charondo 47
4 ChedaJewel Anna Behrens 27461
5 ChedaJewel Ara Avedian 40
6 ChedaJewel Barbara Inwald 164
7 ChedaJewel Brittany Livingston 770

Image uncertainty by cataloger:

Absolutely sure No Response Not sure Pretty sure
ALEX WAGSTAFF 886 6 0 0
ALEXANDRA MATTHEWS 337 27 0 0
ALICE WU 9 12 0 0
ALICIA BLOSE 181 0 0 0
ALIYA MCCARTHY 336 9 0 0
ALYSSA GHIRINGHELLI 421 0 0 0
AMANDA MAGALLANES 321 0 2 0

Image uncertainty by cataloguer (percentage):

Absolutely sure No Response Not sure Pretty sure
ALEX WAGSTAFF 886 6 0 0
ALEXANDRA MATTHEWS 337 27 0 0
ALICE WU 9 12 0 0
ALICIA BLOSE 181 0 0 0
ALIYA MCCARTHY 336 9 0 0
ALYSSA GHIRINGHELLI 421 0 0 0
AMANDA MAGALLANES 321 0 2 0