TAEMILE Project

TAEMILE: Towards Automating Experience Management in Interactive Learning Environments

Playable Demo

Unity web player not available at the moment.

Note: The current builds contain telemetry collection. When the system prompts for an ID, leave the field blank and press "continue".

This software is distributed for free for educational and research purposes. The software is provided "as is," with all faults, defects and errors, and without warranty of any kind. We cannot warrant that the software will be free of bugs, errors, viruses or other defects, and we shall have no liability of any kind for the use of or inability to use the software, the software content or any associated service. Telemetry will be collected and sent to a data collection server. The collected data is limited to the in-game behavior and screen resolution. By downloading and executing this application you acknowledge this notice and permit the collection of this data. This software contains code by third parties. Built with Unity. All assets are copyrighted 2014-2016 Drexel University. Trademarks and brands are the property of their respective owners.

Dataset

A limited version may be available to select parties, contact us for more information.

Additional information

Dr. Jichen Zhu projects page
NSF press note
Gameplay footage
Play styles footage
Wizard of Oz interface demo (early prototype)

Publications

This is a list of some of the project's publications.

J. Valls-Vargas, A. Kahl, J. Patterson, G. Muschio, A. Foster, J. Zhu (2015). Designing and Tracking Play Styles in Solving the Incognitum. GLS 2015. [PDF pp. 241-247]
- @inproceedings{valls2015gls,
  - Author = {Valls-Vargas, Josep and Kahl, Andrew and Patterson, Justin and Muschio, Glen and Foster, Aroutis and Zhu, Jichen},
    - Booktitle = {Proceedings of the Games + Learning + Society Conference},
    - Date-Modified = {2015-07-01 21:12:07 +0000},
    - Keywords = {educational games, player modeling, ILE},
    - Organization = {Games + Learning + Society},
    - Pages = {to appear},
    - Title = {Designing and Tracking Play Styles in Solving the Incognitum},
    - Year = {2015},
  - }

Technical information

The game environment has been designed to record game events in a log file and transmit game telemetry. The telemetry data captures movement and interaction variables and can be used to recreate the play session.

A copy of the log files will be stored in your system's persistentDataPath.

The events recorded are:

Close: close a card
End: finish the game
Focus: open a card
HelpCardShown: a tutorial item is shown on screen
InfoBox: the card is visible
InventoryNew: they visited a new exhibit
KeyDown
KeyUp
Look: looked at an object in the environment (not only exhibits)
MapHoover
MapHooverOut
MastodonAll
MastodonNew: they got a new bone
MouseDown
MouseMoveD: mouse movement
MouseMoveS: mouse movement
MouseMoveV: mouse movement
MouseUp
Navigate: they started walking
PauseMenu: open/close the concept map
QuestionAsk: a question is asked
QuestionCardAnswerFalse: they answered wrong
QuestionCardAnswerTrue: they answered correctly
Track: activity, position, camera rotation and heading every 500ms

Some of the events have additional parameters, here are some examples:

QuestionCardAnswerTrue/Jefferson: the current card
MastodonNew/Portrait: the type of exhibit
MouseMoveV/0.5640188: movement speed
PauseMenu/Pause: opening the concept map
HelpCardShown/6: the sixth card of the tutorial is shown
MouseDown/0: the left mouse button has been pressed
Close/MastodonBone4: the current card
MouseUp/1: the right mouse button has been depressed
PauseMenu/Unpause: closing the concept map

The log files are segmented into chunks using different strategies and each chunk is transformed into a feature vector with the following features:

# of transitions between quest and optional lines (by exhibits belonging to lines)
# of lines visited (by exhibits belonging to lines)
minimum amount of time spent reading a card
maximum amount of time spent reading a card
average amount of time spent reading cards
minimum amount of time spent reading a question
maximum amount of time spent reading a question
average amount of time spent reading questions
average distance from beacon located in the NW corner of the room
average distance from beacon located in the SE corner of the room
average distance from beacon located in the NE corner of the room
distance walked (M)
average walking speed (M/X)
time of the current segment (X)
time spent in cards (reading and answering questions) (A)
time spent checking the concept map (quest tracking tool) (B)
time spent navigating the environment (not in cards nor in concept map) (C)
ratio of time spent in cards (A/X)
ratio of time spent checking the concept map (B/X)
ratio of time spent navigating (C/X)
# of times the concept map (quest tracking tool) has been opened (M)
# of items hovered with the mouse while in the concept map (N)
# of times an item in the quest lines has been hovered (R)
ratio of times hovering a quest item and non-quest items (R/N)
# of different items hovered
# of items hovered for less than 500ms
average time an item is hovered on with the mouse (T/(R-S))
average # of items hovered each time the concept map is opened (N/M)
average time spent at the concept map when opened (B/M)
# of times concept map opened was opened within 5 seconds
# of key presses (down) (O)
# of mouse button presses (down) (P)
# of events recorded (Q)
key presses per second (O/X)
mouse button presses per second (P/X)
events recorded per second (Q/X)
# of questions answered correctly (D)
# of questions answered incorrectly (E)
# of questions answered correctly in a row
ratio of questions answered correctly (D/(D+E))
ratio of questions answered incorrectly (E/(D+E))
# of words shown in cards
average reading speed for each visit in a card
average reading speed for visits and revisits in cards (i.e. revisiting an exhibit increases the average)
# of words shown in questions
average reading speed for each visit in a question
average reading speed for visits and revisits in questions (i.e. revisiting an exhibit increases the average)
# of new items visited (F)
# of items revisited (G)
# of visits (F+G)
# of new questions visited (H)
# of questions revisited (I)
# of questions (H+I)
time spent in exhibits completing a quest (J)
time spent in exhibits in quest lines (K)
time spent in exhibits in optional lines (L)
ratio of time spent in exhibits that complete a quest (J/A)
ratio of time spent in exhibits in quest lines (K/A)
whether one of 10 specific patterns of interaction (sequences of 3 to 7 actions) happened (e.g., open concept map, walk, interact with an exhibit on a quest line, open the concept map)
some additional identification related to the session (e.g., telemetry ID) and gameplay (e.g., screen resolution, date/time started)