Making an analogy is a common mechanism we use to create a sense of shared understanding. In the baseball scouting community, this is standard practice, especially when communicating the physical attributes or abilities of a specific player. The problem is that these player comparisons are largely dependent on our personal encyclopedia of players, and are rarely able to marry the physical attributes with other ability characteristics or production capacity. ScoutEm automates the comparison process by auto-suggesting similar pitchers that match on both physical attributes and pitch characteristics.

What it does

ScoutEm allows the user to select a target pitcher, and then suggests comparable pitchers by matching on physical attributes, pitch metrics, and production statistics. Our proprietary optimization algorithm best fits a list of analogous pitchers based on predetermined selection criteria. No longer are we limited by a narrow player library, or reaching to make comparisons between to pitchers.

How we built it

We built Scout Em using Node Js, React Js and Bootstrap for styling.

Challenges we ran into

We were able to pull most of the data needed for this project from the SportRadar API. One of the early challenges was that the API only allows you to query the database on an ID number, and not last name. We hardcoded in a few test players, but would need to build in this functionality.

Accomplishments that we're proud of

Solves real world problem!

What we learned

Team work and present .

What's next for ScoutEm

What we've presented today is a very preliminary model. ScoutEm would ideally allow the user to select the fields which are most defining of a specific player and manually enter the parameters of a hypothetical player. This application is most useful for comparing players across leagues, i.e. an amateur pitcher to a Major Leaguer. We would need to develop the similarity algorithms, and expand beyond just pitchers to all players. The addition of video and links to other public databases are logical next steps.

Share this project: