This is the heroku deployment of Kick-Into-Traction (KIT). For modelling repository please check: https://github.com/nmosto/kick-into-traction
Kick-Into-Traction (KIT) is a web app that makes recommendations to users of Kickstarter's crowdsourcing platform on how to improve the likelihood of a campaign's success by optimizing features related to the images of a campaign.
KIT has analyzed tens of thousands of Kickstarter campaigns to try and uncover what image features differentiate successful campaigs from the unsuccessful ones.
Some Key Features:
- Dullness
- Brightness
- Edge Behavior
- Blurriness (variance of the Laplacian filter)
- Average RGB
- Dominant Colors (with K-means Clustering)
- Compression Size
- Aspect Ratio
Coming Soon:
- Simple Object Identification (convolutional neural networks)
Web app found here: https://kit-streamlit-heroku.herokuapp.com/
More information coming soon!