Automatic Analysis of Mobile App Reviews


Universal Bank operating in financial services through its 120 branches and 877 ATMs and Self-Service Terminals in 17 provinces of Argentina. Their staff of   3700 employees assist   1,000,000 homes, 40,700 Pymes and commerces/shops and approx.  1,821 corporate customers. One of the market leaders in secured loans, electronic banking, leasing and foreign trade. Their assets extends to over 4,900 millon dollars while their deposits are around 2,500 millon dollars. They rank between the 7 first private banks in Argentina.


The bank needed the automatic analysis of the reviews of their mobile application in order to construct the problems found by users and objectively monitor the evolution of the impact subsequent to   the correction actions with the intention to enhance their users satisfaction and fidelization.


The qualification and reviews of applications in platforms like Google Play Store deliver valuable information on the client experience becoming thus a very important tool to understand what their likes and dislikes are regarding the product/service offered through the app. Thanks to machine learning we can analyse myriads of reviews of this type and know where the product/service/application does not meet the client’s expectation. Through the use of Natural Languaje Processing (Topic Modeling) algorithm all reviews are automatically scanned and all different topics are then labelled as a function of time. These topics then feed the developer’s   backlog to enhance the identified app aspects.


  • Visibility and traceability of the application issues.
  • Prioritization of solutions as per type of issue detected.
  • Enhanced user experience thanks to improvements implemented on the app.