// Inspiration These brazen and often confusing times have often led to the most bizzare conversations pertaining to what if scenarios. Even media agencies, news channels have delivered information in the most bemusing ways. The narrative keeps changing almost daily, and all our thoughts have become even more so staggered. This prompted us as a team, to go ahead with making of an app, that would help us organise our randomest thougts, the most vital of information, any reminders that we might have, any interesting anecdote, that we just remembered, because of the never ending deja-vus of our childhood, post having to come back, and stay put with our family. In this day and age, using a pen and paper to do so, just feels too cumbersome, and all the services in the market are either too restrictive in nature, or generally are clouded with the doubts of absolute privacy. As soon as we saw the potential of what wit.ai can do with a small training set, It was clear in everybody's mind in our team, that a smarter journal is what we were going to be making.
// What it does JournAI seems like any other mundane note keeping app in the market, capable of storing journal entries both via voice and text forms. However the lucidity of leveraging the power of wit.ai 's API to get the meaning of sentences at a coarse level helped us propel our application to something more. JournAI breaks down the meaning of the note stored by the user, all of them unassorted, but it provides the user with a very powerful fuzzy search, ranking the notes most relevant to query made by the user. The USP of the product lies in the fact, that besides accessing the usual way in which notes are displayed according to the time of creation, relevant notes are found, and the sorting of the notes are done pretty quickly even if the note contents are incosistent in their turn.
// How we build it The application was built as an Android Application, made using Kotlin as the base language, supporting SDK 23 and higher. The database used for the backend, is stored locally, since the expected size of the text is minimal, and due to the need for privacy, the entire database is stored locally in the private folder of the app.
// Challenges we ran into We took part in this competition to learn something new, and with this vision, we decided to not go ahead with tools that we were already familiar with. No one in the team new how to code in Kotlin, and that served as the gateway to many a challenges with regards to coding in such a powerful, albeit comparatively less documented language for Android application development. Another major hurdle was the never ending debate of what to keep and what not to keep, amongst the user defined entities, and possible intents of each of the sentences uttered. Besides, journal entries can be arbitrarily long, whilst the restriction on the API query made to wit.ai is to not allow more than 280 characters in a single go. This disparity created by differently sized journal entries, made the designing of the scheme to store all entries tougher.
// Accomplishments we are proud of Not so much of an accomplishment, but this application is one of the first, that actually got made in the way, that we intended it to be, and is therefore something that will come in handy for us in our day to day proceedings, something we hope follows suit, with people who come across this application.
// What we learned This entire experience has been rewarding, we learned a lot right from an introduction to an amazingly simple, but extremely powerful platform that can be leveraged to make applications, that would have earlier taken a lot more effort, time and skill, but can now be made with way less effort. Besides, we got to learn a new language, and increased our experience of coding up Android applications, our experience in which was almost negligible.
// What's next for JournAI When we started out, the idea was to make JournAI a Spotify kind of an equivalent for searching across our journal entries. Giving us, some daily thought of the day suggestions, upon gauging the overall mood we are in, inferred from the kind of journal entries we have been writing. Organizing our journal entries in a way, that allows for a cross connection to other apps, that help us set reminders. There is also an intention to expand the application to put in a reminiscince column, showcasing old anecdotes, that were written by the user, again depending on the kind of the mood gauged, by the kind of entries being made. Finally we also wanted to incorporate a Vision component so as to include figures, and have even more powerful search options. We have other disconnected vague future plans as well, and will probably be using JournAI itself to organise all of our plans into a coherent roadmap for improvements.