- upload your proposal to this directory as proposal.pdf
- edit this README file to include your timeline. Fill in the We plan to accomplish section for each week with specific steps you plan to have completed by the lab meeting that week.
Each week, for the checkpoint you need to fill out by Monday Night (so I have a day to read it before lab):
- Which of the goals have you accomplished
- For the goals you did not accomplish, state the issue/problem that is blocking progress
- Any questions you have. This can help us prepare for the lab meeting.
- Update your timeline/goals for the following week if needed
This is what you plan to accomplish in the first week of the project, leading up to our lab meeting. You must fill this in as part of your proposal.
- Hand in the full proposal
- Figure out how to scrape pitch data
- Begin work on World Series data
Complete this section ahead of lab for week 1 of the project
- Completed: Project proposal, timeline, scraped data
- In progress: Working on World Series data
- Stuck: Nothing at the moment
None right now.
This is what you plan to accomplish in the second week of the project, leading up to our lab meeting. You must fill this in as part of your proposal; update after each checkpoint as necessary (i.e. as your plan evolves, this document should reflect the changes).
- Have fastball/curveball classifier on World Series data
- Have fastball/curveball classifier on big data set
- Have fastball/curveball/changeup classifier on World Series data
Complete this section ahead of lab for week 1 of the project
- Completed: All of our goals listed above. We also finished the fastball/curveball/changeup classifier for the big data set, and added a confusion matrix for our data to see what pitches were getting mis-identified.
- In progress: Working on adding more and more pitches.
- Stuck: Nothing at the moment
None right now.
This is what you plan to accomplish in the third week of the project, leading up to our lab meeting. You must fill this in as part of your proposal; update after each checkpoint as necessary (i.e. as your plan evolves, this document should reflect the changes).
- Attempt to have a classifier on every pitch in World Series data
- Test out classifier on big data
- Look to run clustering to find trends
- Run statistical tests on classifier
Complete this section ahead of lab for week 3 of the project
- Completed: We completed the big data with our algorithm. We added some more features to aid in the accuracy. These features were release height and side. The statistical tests don't seem to be significant for our data as accuracy is much more important. Also, trying to see what pitches are misdistinguished and if we can create mini forests to help seperate.
- In progress: Working on mitigating error.
- Stuck: Nothing at the moment
We need advice about/help with ... (optional): -Potentially where to go with error fixing, potentially a random algorithm.
This is what you plan to accomplish in the fourth week of the project, leading up to our lab meeting. You must fill this in as part of your proposal; update after each checkpoint as necessary (i.e. as your plan evolves, this document should reflect the changes).
- Write the paper
- Make presentation
- Completed: Improved error on algorithm to reach our goal of greater than 90% accuracy. Analyzed a potential way to distinguish cutters and sliders. Attempted to group similar pitches together, found that is extremely difficult with KNN.
- In Progress: Analysis of potential reasons similar pitches are hard to cluster. Beginning to write the paper.
- Stuck: Nothing at the moment.
No requests at the moment.
- Our last Lab meeting is April 26th
- Experimental work should ideally complete a week ahead of time so you can focus on writing
- Paper due on last day of classes (April 28th)
- Work on Final presentation, to be delivered during final exam slot