Gas adds up--especially for large transactions! We were tired of waiting for longer-than-predicted transaction times and overpaying for gas based on incorrect predictions.

What it does

The program uses machine learning models to predict the best gas prices for five categories:

  • Safe <30 min
  • Standard <5 min
  • Fast <2min
  • Fastest ~1 Block (<30 sec)

The program also compares gas estimates against ETH Gas Station to show how much better our model forecasts gas prices. :)

How we built it

Built using:

  • SKLearn Machine Learning Models
  • Dash
  • Infura

Challenges we ran into

We ran into issues editing Dash as it was our first time using it with callbacks.

Accomplishments that we're proud of

We are proud of the product we produced and for the capability users have to pay significantly less in gas with our forecasting technology.

What we learned

We learned how to use Dash and

What's next for Squid Station

Integrate into current projects with high transaction volume and/or large data transactions.

Built With

  • dash
  • infura
  • sklearn
Share this project: