Application Development
September 2018 - December 2019
The first part of the project was the creation of the prototype. Espy's goal was to be innovative and stand out amongst other similar applications. Team 8 focused majority of the CS 425 semester to create a machine learning model. There were quite a lot of difficulties in the process. For building the model, we decided to use TensorFlow for ML and for IDE we used Google Colab. For a mobile platform we wanted to develop for both iOS and Android using Xamarin.
In out first Prototype Demo Team Meeting with our advisors on December 4, 2018, our team had successfully implemented a test version of the machine learning algorithm. We had very high hopes that the model could be deployed to a messy data set and be able to recognize noise from actual image pattern. Unfortunately, the model did not fit the requirements of our application and we had to eliminate the entire progress to date. That was a major set back for us because we had put all of our time and effort into making that extremely important piece of the application work seamlessly. It was very devastating for us and we had to sit down and have multiple meetings with different external advisors to salvage what we have.
The decision was to continue with the original idea of having a fashion app but eliminate the use of machine learning and incorporate it after Innovation Day as a future work in progress.
In out first Prototype Demo Team Meeting with our advisors on December 4, 2018, our team had successfully implemented a test version of the machine learning algorithm. We had very high hopes that the model could be deployed to a messy data set and be able to recognize noise from actual image pattern. Unfortunately, the model did not fit the requirements of our application and we had to eliminate the entire progress to date. That was a major set back for us because we had put all of our time and effort into making that extremely important piece of the application work seamlessly. It was very devastating for us and we had to sit down and have multiple meetings with different external advisors to salvage what we have.
The decision was to continue with the original idea of having a fashion app but eliminate the use of machine learning and incorporate it after Innovation Day as a future work in progress.
January 2019- May 2019
In the second semester in CS 426, our team had decided on a new approach for the revision of Espy. The first major change was scraping the idea of machine learning and focusing more on the UI and basic functionality.Another important decision to make was committing to a single mobile platform for development. There were some difficulties from the get go about using Xamarin because of its compatibility with ML and deployment of the application. A final decision was made to develop for iOS using Xcode and Swift, which are proprietary to Apple. We encountered some difficulties with Xcode due to the syntax changing with every Swift version, making it an inconvenience developing the application. Nevertheless, it was a rewarding process because we learned so much on our own through trail and error. One of the really supportive aspects of the development process was Apple's Developer Documentation tools, which were our most important development support.
Currently, Espy is in stable working condition. We have a main menu with a built in camera for the user to take picture and create an inventory of their garments and shoes. There is an outfit viewer tab, where the user can create fun daily outfits or use the random button generator to ask the system for a suggestion.
It was very important for us to make a simple UI for all types of users to be able to take advantage of the application. In the YouTube link at the bottom you can see a quick demo of what Espy is all about and get to know how to use the app and what to expect of the implementation.
Currently, Espy is in stable working condition. We have a main menu with a built in camera for the user to take picture and create an inventory of their garments and shoes. There is an outfit viewer tab, where the user can create fun daily outfits or use the random button generator to ask the system for a suggestion.
It was very important for us to make a simple UI for all types of users to be able to take advantage of the application. In the YouTube link at the bottom you can see a quick demo of what Espy is all about and get to know how to use the app and what to expect of the implementation.