Fantasy FortuneTeller

Fantasy FortuneTeller makes fantasy football easy! FF allows you to compare NFL players and their current weekly fantasy projections without being bombarded with unnecessary stats.

Please visit our production app here.

How Does It Work?

FF makes use of a secondary Sinatra app (link here) that holds current NFL player data and projections from several external APIs (Fantasy Football Nerd, SportsData, and nflscrapR). Within FF, you can easily make teams and add players to those teams, allowing straightforward fantasy projection comparisons for the current NFL week.

Getting Started

• Clone this repository
• CD into this repository
• Run bundle
• Run rake db:{create,migrate,seed}
• There are 4 rake tasks necessary to update the current player data (these tasks do require that you have a key to the FF Sinatra API- please contact one of the application authors for more information)
  • rake import:players to load all current NFL fantasy players into the db
  • rake import:logos to load NFL team logos
  • rake update:injuries to update current player injury statuses
  • rake update:projections to get the current week fantasy projections for each player

Running the tests

To view the testing suite, please run rspec from the root directory /fantasy-fortuneteller

Future Iterations

Originally, we had hoped to complete one or all of the following features for out MVP. However, finding access to data that was both affordable and throrough enough for our purposes turned out to be more challenging than anticipated. Given more time, we could have pulled in enough data from addidtional sources and built a more robust API necessary for the data analysis and machine learning we wanted to build. But given the timeframe of the project, we decided to redefine the scope of our MVP and focus on the following features in future iterations:

• Compare projected vs. performance calculations for the past 3 weeks to determine if a player is hot or cold.
• Calculate confidence intervals from each players historical data to give users the option to see projections as a range.
• Compile and process historical play-by-play data to build a machine learning algorithm to create our own projections.

Behind the Scenes

• Ruby 2.4.1
• Rails 5.2.3
• DB: PostgresQL
• CI: TravisCI
• Testing: RSpec
• External API calls: Faraday
• Background workers: Redis and Sidekiq
• User experience: HTML, CSS, and JavaScript
• Performance metrics: New Relic

Authors

• Fenton Taylor - Fenton Taylor

• Alec Wells - alect47

• Tyler Schaffer - tschaffer1618