Update 1: Our Plan

[seandube]

ESPN plans to start a new project. Instead of focusing on the MLB, we plan to look at NFL statistics. We are very curious about the outcomes of plays, so we want to look at downs, distance to go for the first down, yard line, player position, and play type. We want to see if we can predict the outcomes of plays based on these variables. We also want to look at tracking data to determine the average position of players on the field. The data we plan to explore is given by the NFL's Big Bowl Data competition. It is provided in CSV format which will be fairly easy to load into R.

The final outcome will be a Shiny App. The user will be able to input yard line, downs, distance to go, play type, and the app will return a choropleth of a football field and will either show a heatmap out the outcomes of plays or player icons showing player positions. We would have another tab with the same input, but instead returns the actual numbers like average yards gained, first down percentages, and touchdown percentages for the specific inputs.

In terms of our schedule, by update 2, we plan to have finished data wrangling and start the shiny app. By the third update, we plan to have finished ⅔’s of the shiny app. In the week between update three and the final presentation, we plan to finalize the shiny app and record our presentation.

[katcorr]

Nice plan, team! An exciting new direction. A few questions/clarifications:

• "We want to see if we can predict the outcomes of plays based on these variables." Could you define "outcome of play"? What are the categories for "outcome of play" going to be?
• "We also want to look at tracking data to determine the average position of players on the field." I'm not sure taking an "average" makes sense here? Is that really what you meant (take the mean of coordinates?) or something else?
• to clarify: you do need to create a blog post with written introduction, conclusion, limitations, etc.. You can link to a Shiny app in your blog post, and have that Shiny app contain the meat of your project (visualizations etc.), but you still need to write an engaging and informative blog post.

Could you please update your schedule to incorporate time for the writing of the blog post and respond to my questions above? Please do so by responding to this comment by 11/3.

Update 1: 10/10

[NathanTatko]

Q1: When we say we are trying to predict the outcomes of plays, we are talking about finding the average yards gained for the specified inputs, first down percentage (percentage of plays that result in a first down for the offense), and touchdown percentage (percentage of plays that result in touchdowns for the offense). Using these numbers we could predict which plays will be most successful for the offense.

Q2: For the tracking data, we are going to plot the average starting location for every position and also the average final location of every position for each play based on the specified inputs. This will show us how players tend to move based on the down, distance to go, and the yardline.

Schedule: By update 2, we plan to have finished data wrangling and started the shiny app. By update 3, we plan to have finished first plot of the shiny app and have started the second. In the week between update three and the final presentation, we plan to finish the second plot, write up the blog post, and record our presentation.