19/10/2015 Supporting Developers in Making Successful Apps Federica Sarro University College London
Use the information available in mobile app stores to support developers in making successful apps
Use the information available in mobile app stores to support developers in making successful apps Requirements Elicitation from App Description Characteristics of Impactful App Releases Challenges Make it relevant Make it reliable
Making it Relevant
Making it Relevant Are app owners/developers really interested in these findings? Do they analyse user reviews for the benefit of next releases? Do they monitor the user feedbacks of competing apps? A. Al Subanin, F. Sarro, S. Black, L. Capra, M. Harman
Making it Relevant Are app owners/developers really interested in these findings? Do they analyse user reviews for the benefit of next releases? Do they monitor the user feedbacks of competing apps? What else can be helpful? A. Al Subanin, F. Sarro, S. Black, L. Capra, M. Harman
Making it Relevant Are app owners/developers really interested in these findings? We are interviewing app developers to discuss their experience with app stores To understand their point of view and make it count in the upcoming exciting research in app stores If you are an app developer interested in the interview, please drop an email to us: afnan.alsubaihin.14@ucl.ac.uk, f.sarro@ucl.ac.uk A. Al Subanin, F. Sarro, S. Black, L. Capra, M. Harman
Have You Ever Mined Data From Mobile App Stores? W. Martin, M. Harman, Y. Jia, F. Sarro, Y. Zhang
The App Sampling Problem for App Store Analysis by W. Martin, M. Harman, Y. Jia, F. Sarro, Y. Zheng, MSR 15 Effect of enforced sampling bias Sample size or content is outside of our control App Store data availability Memory or framework limitations W. Martin, M. Harman, Y. Jia, F. Sarro, Y. Zhang
The App Sampling Problem for App Store Analysis by W. Martin, M. Harman, Y. Jia, F. Sarro, Y. Zheng, MSR 15 Authors App Store Set of Set of Type of Reviews Apps Dataset Hoon et al. (tech report '13) Apple 8 700 000 17 000 F Iacob and Harrison (MSR'13) Google 137 000 270 Pa Carreño and Winbladh (ICSE'13) Google 327 3 Pa Khalid (ICSE'13) Apple 6 390 20 Pa Fu et al. (KDD'13) Google 13 286 706 171 493 F Pagano and Maalej (RE'13) Apple 1 100 000 1 100 Pa Iacob et al. (BCS-HCI'13) Google 3 279 161 Pa Chen et al. (ICSE'14) Google 241 656 4 Pa Guzman and Maalej (RE'14) Apple & Google 32 210 7 Pa W. Martin, M. Harman, Y. Jia, F. Sarro, Y. Zhang
Impactful Mobile App Releases Characteristics of releases that achieve high impact release text length release day rating price new features bug fixes number of ratings download rank W. Martin, F. Sarro, M. Harman
Impactful Mobile App Releases 1,547 releases (Google Play and Windows Phone) 40% (G) and 55% (W) of releases impacted app performance higher prices, day of release and fewer mentions of bug fixing can increase the chance for a release to be impactful higher prices, more descriptive release text and new features can increase the chance for a release to improve rating W. Martin, F. Sarro, M. Harman
Feature Lifecycles as They Spread, Migrate, Remain, and Die in App Stores Federica Sarro, Afnan Al Subaihin, Mark Harman, Yue Jia, William Martin, Yuanyuan Zhang University College London UCL App Store Analysis http://www0.cs.ucl.ac.uk/staff/f.sarro/projects/uclappa/home.html
App Store: A New Avenue for Software Requirements Users Feedback
App Store: A New Avenue for Software Requirements Users Feedback
App Store: A New Avenue for Software Requirements Users Feedback
App Store: A New Avenue for Software Requirements Users Technical Feedback Business Developer
App Store: A New Avenue for Software Requirements Users Technical Feedback Business Capturing user reactions Identify trends across the app store Developer
Migratory Behaviour of App Features Across Product Categories List Event
Migratory Behaviour of App Features Across Product Categories List Event Set Theoretic Characterisation Empirical Study
What Developers May Ask Which migratory behaviours carry monetary value? Which migratory behaviours involve more popular features? Which categories are more likely to migrate features to one other?
What Developers May Ask Which migratory behaviours involve more popular features? Points Of Interest List Events Show Contact Detail Email Picture
Migratory Behaviour of App Features Across Product Categories List Event Set Theoretic Characterisation Empirical Study
Set Theoretic Characterisation of App Store Feature Migration Migratory Behaviour Non- Migratory Behaviour The Theoretical Feature Migration Subsumption Hierarchy
Snapshots App Database snapshot t0 snapshot t1 snapshot t2 snapshot t3
Snapshots App Database snapshot snapshot t1 snapshot snapshot Category 1 F1 Category 2 Category Membership F2 F1 F3 Category 3 F3 F4 F1 F3 is member of is member of { { { {
Weak Migration F C1 A feature weakly migrates if it resides in at least one new category at the end of the time period considered snapshot t0
Weak Migration F C1 C2 A feature weakly migrates if it resides in at least one new category at the end of the time period considered snapshot t0 snapshot t1
Strong Migration F C1 C1 C2 A feature spreads from at least one category to at least one new category and remains in all categories in which it originated snapshot t0 snapshot t1
Weak Exodus F C1 C1 C2 C3 A feature disappears from at least one of the categories in which it previously resided, while appearing in at least one new category snapshot t0 snapshot t1
Strong Exodus F C1 C3 C2 C4 A feature disappears from all categories in which it previously resided to take up residence in at least one new category snapshot t0 snapshot t1
Intransitive F C1 C1 C2 C2 An intransitive feature neither appears in any new categories nor does it disappear from any between the start and the end of the time period considered snapshot t0 snapshot t1
Intransitive F C1 C1 C2 C2 An intransitive feature neither appears in any new categories nor does it disappear from any between the start and the end of the time period considered snapshot t0 snapshot t1
Weak Extinction F C1 C1 C2 A feature disappears from at least one category in which it resided and does not migrate to any new ones snapshot t0 snapshot t1
Weak Extinction F C1 C1 C2 A feature disappears from at least one category in which it resided and does not migrate to any new ones snapshot t0 snapshot t1
Strong Extinction F C1 C2 A feature disappears from all categories in which it resided and does not migrate to any new ones snapshot t0 snapshot t1
Migratory Behaviour of App Features Across Product Categories List Event Set Theoretic Characterisation Empirical Study
Subjects Weeks 3 and 36 in 2011 1,324 non-free features (w3) Weeks 5 and 36 in 2011 623 non-free features (w5)
Subjects Weeks 3 and 36 in 2011 1,324 non-free features (w3) Weeks 5 and 36 in 2011 623 non-free features (w5) No App Sampling Problem Martin et al. MSR 15
Extracting Features from Description of Apps Samsung App Store Mark Harman, Yue Jia, Yuanyuan Zhang. App Store Mining and Analysis: MSR for App Stores. MSR 2012: 108-111
RQ1: Feature Migration How do features distribute over the different migratory behaviours?
RQ1: Feature Migration Weak Extinction Strong Migration Intransitive Weak Exodus Weak Extinction Strong Migration Intransitive Weak Exodus
RQ1: Feature Migration Features often migrate to a category that has similar characteristics Find Location Maps & Navigation Travel
RQ1: Feature Migration Some features have clearly transferable value allowing them to migrate across category boundaries LATEST NEWS News Sport & Recreation
RQ1: Feature Migration Intransitive features are features that are crucial to a given category [view, gps, status] Navigation [sort, track] Music/Video
Which Feature Shall I Include in My App? Intransitive features > must-have requirements for apps in a given category Early identification of the migratory features > undiscovered requirements
Other Research Questions RQ2: Differences in Migratory Behaviours Are there any differences in the price, rating and popularity of features that exhibit different migratory behaviours? RQ3: Correlations among Price, Rating and Popularity Are there any differences in the correlations between price, rating, popularity within each form of migratory behaviour?
Other Research Questions RQ2: Differences in Migratory Behaviours Are there any differences in the price, rating and popularity of features that exhibit different migratory behaviours? RQ3: Correlations among Price, Rating and Popularity Are there any differences in the correlations between price, rating, popularity within each form of migratory behaviour?
Future Directions Interplay between feature migration and user requests
Future Directions Interplay between feature migration and user requests Relationship between migrations of claims and migrations of code
Future Directions Interplay between feature migration and user requests Relationship between migrations of claims and migrations of code Feature migration in other kinds of software system (e.g. software product lines)
Future Directions Interplay between feature migration and user requests Relationship between migrations of claims and migrations of code Feature migration in other kinds of software system (e.g. software product lines) Other longitudinal studies of app stores over periods of time
f.sarro@ucl.ac.uk
UCLappA website http://www0.cs.ucl.ac.uk/staff/f.sarro/projects/uclappa/home.html f.sarro@ucl.ac.uk