Predicting your walk-up song using your facial expression

• The experience: We created an application that takes a photo of someone's face. The physical attributes are then analyzed to determine the emotional expression. This analysis is based on facial attributes such as the position of the eyes, nose, and corners of the mouth. A song then plays based on the corresponding emotion.
• How it works: The facial image is analyzed by Amazon Rekognition to determine the expression. Once the expression is automatically detected, the application connects with Spotify to select a track and then "casts" to a custom Spotify player application (e.g. speakers). The photo is discarded, and the emotional expression is logged for future statistical analysis. For example, we can tell how many "angry" faces are logged during each session, but not which faces are angry.
• Where else you might encounter this technology at work: Imagine this feature helping a car detect when its driver is drowsy, or a concert venue or theater detecting a crowd's reaction to a product or performance.

Gathering real-time stats for a personalized baseball card

• The experience: As attendees participate at different stations during the Batting Practice event, their individual stats and photo are logged by volunteers using an application. Once a card is completed, the data is pushed to a temporary data table, which a design application uses to automatically construct and publish the baseball card for printing.
• How it works: The data and images are managed by a central database. The player's completed data is sent to the temporary data store when the baseball card is ready to be published. The captured data is logged anonymously for future statistical analysis.
• Where else you might encounter this technology at work: This basic structure could be used to collect any type of unstructured or semi-structured data. Many survey applications work similarly, allowing respondents to answer the questions they want and aggregating the data to anonymize the results.

Driving competition through a live leaderboard

• The experience: As data is collected from the various batting practice stations, the leaders in each of the three categories of data (pitching, hitting, and running) are updated and displayed on the video board and TVs around the stadium. The leaderboard rotates through the different categories on a timed basis. Attendees can also see how the whole group is doing with live statistics and challenging targets that drive everyone’s participation.
• How it works: The platform pulls real-time data on demand from the central player management database and organizes the leaderboard rankings by category and "league" (or, whether the participant is a child or adult). We also analyze the aggregated stats of all attendees and compare against real historical player achievements in real-time. We also analyze the aggregated stats of all attendees and compares against real historical achievements in real-time.
• Where else you might encounter this technology at work: This concept could be used for showing travel times on interstates or within airports, or anywhere data is collected in real-time between two points and periodically queried. An example is understanding how long it will take to walk to your gate with current crowding when traveling through a terminal or using a collaborative leaderboard encouraging a call center to achieve better customer success.

Personalizing the post-event attendee experience through a custom "Scouting Report"

• The experience: Following the event, each participant receives a "Scouting Report" that summarizes the player statistics throughout the event. The Scouting Report provides the closest match-up to a Chicago Cubs player and determines a player archetype (e.g. "power hitter," "pinch runner," "pitching ace," etc.)
• How it works: We extract the same data stored in the central player management database and advanced analytical detail from sensor-connected devices and feed it into a database. A data science/machine learning model is run to clean the data and develop various possibilities and analytical comparisons. The resulting output is then fed into a visualization tool to build the Scouting Reports.
• Where else you might encounter this technology at work: Data science is used for customer segmentation, image recognition, insurance fraud detection, delivery store optimization, and more. For example, the more (and better) data you have on your customers, the better you can analyze and determine trends to inform how you reach those customers and ultimately grow market share.