WorkPod

A collaborative platform for project development with a unified interactive dashboard for progress updates and a personalized song recommender based on your mood as music therapy!

Comment

Inspiration

So WorkPod is based on a problem statement that I've come across numerous times especially during college project submissions and I'm sure you have too. You ever had an idea for a project, but you weren't sure what steps are required from you to be able to complete it? Or maybe you formed a team and nobody is certain about what task is being done by which member and how much progress has actually been made thus far? Sometimes, a few members are pushing deadlines or starting to work much later than others? If these issues seem relatable, then WorkPod is the solution for you!

What it does

WorkPod is a platform that keeps the user in mind during each stage of project development. You simply create an account with a project id that can be shared with your team members. Then you share the premise of your project with Arctic and it will break down your idea into a list of tasks with their own estimated timeframes for completion. This list is pushed to OneDash, an interactive project dashboard that is visible to all registered project members, wherein you can see progress of completion of each task as well as user contribution ratios. For each task, you can either click on Complete once you are done to update the dashboard charts or Delete if you do not wish to pursue that task during your project development. I've also created a music therapy app within, so if you head on to Oasis, you can select your mood from the given choices and Arctic, along with data extracted using Spotify's API, will recommend certain songs that will best fit your mood and/or possibly uplift it!

Who created it

SanskarJadhavpic

I am Sanskar Jadhav, a 20 year old third year undergrad in Symbiosis Institute of Technology, Pune, India. I am currently pursuing a B.Tech degree in Artificial Intelligence and Machine Learning, with an Honour's in Aerial Robotics and Drone Technology. I have a passion for web development as well as model training and tuning. I have worked with Streamlit before and am also skilled with React JS and Node Express. My strengths include Statistics, Python, C/C++, and SQL/NoSQL.

How I built it

I built WorkPod on Streamlit using SQLite tables for users and for tasks. Users save their project id, username, email, and profile picture. Tasks are saved after being filtered from Arctic's response to your project idea, with the respective project id and a Boolean value for completion as well as the username for who completed it. I've used Plotly for the charts, specifically a Pie chart for user contributions ratio and a progress bar for project completion status. I created all logos using free online makers, and the Spotify music dataset is roughly 26,000 songs collected using the Spotify API, so that by adding track_id to a common beginning URL, I could create clickable links to direct the user to each track's URL on Spotify. The recommendations are made specifically for each mood based on danceability, energy, acousticness, speechiness, valence, and tempo, with the help of Arctic.

Challenges I ran into

There were quite a few at first! I had to go through the documentation and code for Arctic Instruct in depth. Figuring out how to add Arctic to my Streamlit page was the first proper challenge. Especially since I was planning on tuning the prompts and in my music recommendation page, I was even performing an extensive filter on the expected response. Figuring out which temperature value was optimum for the music recommendation required some hands-on experimenting, and ensuring that proper error handling was developed in case the response did not match my filter's expectation. Also, pushing the tasks from one page to another, since WorkPod is a multi-page app, was another prominent challenge. I aimed to create a push button after the tasks were generated, but that led to errors in accurate task uploading to the SQLite table, so I left it automated and instead incorporated a Delete Task function.

Accomplishments that I'm proud of

I'm really proud of the final product! I think this app could definitely extend well beyond a demo and actually be used by team members across the globe. I'm very happy with the way the UI has been designed and I'm glad that by using my own music data, I am able to incorporate Hindi, Tamil, and Telegu songs, that will also be recommended alongside English songs, since there is no demographic region influence in my recommendation algorithm (as music is music, irrespective of language!), so you can be introduced to famous Bollywood and Tollywood songs from India!

What I learned

Arctic Instruct is a very interesting LLM chatbot, and I noticed that a temperature value beyond 2 would give some interesting results! It definitely has a lot of potential, and I've also gained a deeper appreciation for the ease of developing websites using Streamlit.

What's next for WorkPod

I plan to incorporate account protection (for now, I've only kept username and email so it's easier for me to remember all the accounts I made during testing), as well as increase the music dataset size to include songs from all over the world. I would also like to increase connectivity between team members within the website.

Built With

Share this project:

Updates