Published: Jan. 4, 2022

During our summer, we formed our cross-strand Fall 2021 Data Collection team to begin the all-important task of going into real K–12 classrooms to support teachers implementing our Curriculum Units and collect the data we need to develop our AI Partner. 

This fall, with the help of 24 volunteers, researchers, and iSAT staff members, we collected 352 audio sessions with approximately 24 hours of student collaboration data, and about 150 video files of students collaborating on the unit. 

Middle School students working on a plant sensor.

Denver Public School students measure the amount of water present in this plant's soil using a sensor.

3 St. Vrain Valley Schools (SVVSD) and 9 Denver Public Schools (DPS) teachers implemented the 5 lesson curriculum across 13 classrooms and over 39 class periods during the fall while iSAT researchers observed the classrooms and collected video and audio data of students working together. Data was only collected from consenting students (and their parents), and their interactions were recorded with a Yeti microphone connected to an iPad as our main setup. We also collected comparison audio from a Sony recorder, or a Zoom handheld recorder. A camcorder and a wide-angle camcorder captured video of the consented groups. Sometimes an iPad was used to record close-up videos of interactions among students.

Middle School students work on the sensor immersion unit.

St. Vrain Valley School District middle school students work on the sensor immersion unit.


As data comes in a team at the back-end is hard at work at cataloging it, checking quality, doing preliminary annotations and transcriptions, and making recommendations for the next round. So it is very much a continuous process and we are now gearing up for a new round of in-classroom data collection with a new cohort of teachers starting this Spring.