Inspiration

In a modern world, content discovery is broken. Ok, not broken but twisted. Every day millions of people trying to find some interesting or unusual content that suits their interests, but it's very difficult to broke through the marketing bubble.

What it does

Our goal is to show all available options on one screen in casual datamining manner (Analyze First. Show the Important, Zoom and Filter, Analyze Further, Details-on-Demand)

How we built it

Tech stack we've used:

Client

  • WebGL 1.0 via ThreeJs for rendering
  • C++ via Emscripten for performance processing
  • Semantic UI

Server

  • Python && C++ for data processing and ML
  • SQLite as local data storage :)
  • Nginx for static data hosting

Challenges we ran into

Performance on large datasets was the main bottleneck for all parts

Accomplishments that we are proud of

We achieved reasonable performance and user experience in a short time

What we learned

Naive approaches to architecture do not work on the large datasets :)

What's next for EntireThing

We are planning to do some customer development, testing our UI on the different angles

Share this project:
×

Updates