Nexus - Networking with Data-Driven Surveys | Appwrite Hashnode Hackathon
Command Palette
Search for a command to run...
Search for a command to run...
Hey readers 👋,
Here is my entry for the Appwrite Hashnode hackathon; I hope you like what I was able to create over these two weeks.
Github - https://github.com/manavendrasen/nexus
What are we trying to solve?
Seeking like-minded individuals in conferences and communities can often resemble searching for a needle in a haystack. The vast number of participants makes it impossible to engage with everyone and identify the right connections.
Let's be real, It is not possible to talk to everyone and build your network. If only there was a way which would give me information about people who are similar or dissimilar to me and give me a starting point to build the connection.
Introducing Nexus, a solution that empowers networking in conferences and communities. Leveraging data from a simple survey, Nexus generates an interactive social graph, visually depicting all attendees. Within this graph, individuals with similar interests and characteristics are positioned closer together, facilitating effortless identification of potential connections.
Sounds interesting right? Nexus is just that!! 🤗
A short 3 min video to showcase the product.
Organizers/Community leads' POV
In 3 easy steps, communities and conferences can be superpowered!
They add questions to the survey
Three fundamental questions that act as an introduction are included in each survey.
The survey can have objective questions added by the organisers; the visualisation needs at least three objective questions, but the more the better.
If necessary, they can alter the questions.
Finally, they hit the Publish button to make the survey live!
The generated link can be copied and shared.
Participants' POV
Visit the survey link/code shared by the organizers
Fill out their email and complete the simple form, the progress bar at the top shows the progress.
When the survey is finished, the participants can view the results on the survey link; the logged-in participant is indicated by a green dot.
From here, Participants can see other participants who are similar to them and get the link to connect with them and network.
Organizers can create surveys to understand and interact with the community/conference attendees.
Participants can network with like-minded people to collaborate and discuss.
The app is light mode and dark mode compatible and mobile responsive :)
Nexus empowers communities by enhancing the networking and collaboration potential by using data visualization-driven surveys.
NextJS 13 (Typescript)
Zustand
Tailwind CSS
react-hook-form
Chart.js
AWS Lambda and API Gateway (Python)
SkLearn T-SNE - Dimensionality Reduction while keeping similarities
Appwrite - Auth, Database
How Appwrite helped us?
Appwrite Cloud serverless Auth (Email and Password) and Database made it easy and fast to build the application. Used the client-side SDK to connect to Appwrite directly from NextJS.
How do I visualize the data?
I had no idea how to convert the survey responses into a graph. Research led me to Dimensionality Reduction - PCA and T-SNE. Other algorithms can be used as well, such as Cosine similarity etc.
Deploying the Python script.
Initially, I planned on using Appwrite Functions to deploy the Python script as a simple serverless function. But, my script uses Numpy and Sklearn which are not supported by the runtime Appwrite uses. I would request the Appwrite team to allow custom container images to run the serverless functions. (known issue)
I went with AWS Lambda with a custom image on ECR.
Improve the visualisation algorithm and incorporate text responses into the analysis.
Add more charts like - Most common responses from participants to improve analysis.
PRs are always welcome. If you would like to add some components that you think would be useful while writing docs, you can add/contribute here.
Github - https://github.com/manavendrasen/nexus
Demo - Nexus - Networking with Data-Driven Surveys | Demo
T-SNE - https://towardsdatascience.com
Inspired by - Anson Yu on Twitter
Thank you! 🤍
