Topic Space: a mobile augmented reality

recommendation appJim Hahn and Ben Ryckman

Outline• Funded by IMLS• Prototype Functionality– Native app (Android)– Web app (backbone.js)

• User Study Results• Interface Revision, new functionality

2

IMLS Grant Funding• Sparks Grant 2012-2014• The augmented reality functionality explored

in this grant utilized optical character recognition (OCR) software embedded into a native Android app.

3

Functionality• It allows users to take a picture of a library call

number on a book in the Undergraduate Library book stacks, which then provides recommendations based on the subject area of the call number.

• The app can show users books that are normally shelved in the starting location but that are currently checked out.

4

Modular APIs

5

Webservice for Android and web apps

6

Download Minrva app from Google Play and select Undergrad

7

Native App

8

Native App

9

Native App

10

Native App

11

Web app• Repurposed use of recommender webservice– The recommender references the current item the

student is viewing

12

13

14

15

User Study of Android appWe collected formative evaluation of what students liked about the app, what they found confusing, and if they felt the app would be worthwhile to develop.

16

User Study ResultsStudents went through the bookstacks to go near subject areas where they had courses or to find books that they had previously read, since they wanted to use the app to find other related books that were similar to those they had enjoyed in the past.

17

User Study ResultsIn our debriefing interviews we learned that some students would make use of the recommendation features and other topic areas that the app suggested.

18

User Study ResultsStudents increasingly expect their digital experiences to be similar to those of Google. For example, students expected that the app would show them the location of the phone as it moves through the bookstacks; similar in nature to Google maps functionality. This type of indoor positioning is feasible to prototype

19

User Study ResultsThe biggest hurdle for students in utilizing this app was their expectations for the camera phone; students expected the OCR process to be instant. In practice, we found that the app takes a few seconds to focus and then send then string of captured text to the web server; the web server then takes an additional couple seconds.

20

User Study ResultsAs a result in our revision we looked to replace the input mechanism, moving toward more instant scanning with a barcode scan input, which is much faster, and provides more reliable metadata over which to suggest items.

21

Revisions:New Interface, New Functionality

• User Friendly Design• Easier and More Efficient Input• Superior Suggestion Result Sets• Supplemental Book Information• Easily Request Materials• Easier to Read Mapping Functionality

22

Revisions: Cont.

23

• Simplified Design

• Input Improvements

Revisions: Cont.

24

• Simplified Design

• Improved Result Set

• Book Info and Request

Revisions: Cont.

25

• Easy to interpret mapping

• Pinch Zoom Functionality

Next Steps• Include Functionality for Dewey

Call Numbers

26

Next Steps: Cont.• Add Compatibility to Use Topic

Space in Additional Libraries• Incorporate Cross-Library

Results

27

Next Steps: Cross-Library Results

28