May 4, 2016 | Written by: Andrew Lohr
Categorized: Customer Stories | Watson
Share this post:
Leveraging IBM’s Watson to enhance the way in which educators perform daily tasks is a major focus for an application like Sentry. Sentry is a demonstration web app hosted on Bluemix and built to help elementary through high school teachers better understand how their students feel. Sentry does this by using Watson to identify a learner’s overall emotions and attitude, based on the individual’s written compositions. By analyzing a student’s journal entries or essays, Sentry can alert educators as to whether a student appears to be exhibiting negative, dangerous, or abnormal characteristics, in order to prevent or recognize a possible problem before it develops.
By using Sentry, educators have the ability to create a profile for each of their students. This profile will show every learner’s average personality and tone traits for all papers that have been uploaded. An overall percentage is also continuously calculated based on all entries that have been handed in for a given learner. The traits are dictated through the help of IBM’s Watson Personality Insights, Tone Analyzer, and Document Conversion services.
Sentry allows the teacher to set up sentries on a certain trait with a set threshold. For example, if a teacher were to set up a sentry for the “anger” trait with a threshold of 85%, then Sentry will alert the teacher whenever a student’s average anger trait breaches the threshold. This can allow educators to monitor a student’s mood or overall feelings for a period of time, and intervene when necessary.
In addition to alerting teachers about potentially problematic traits, Sentry can also be used to alert educators about positive traits. For example, educators can create incentives or provide rewards to students who appear to be exhibiting specific positive characteristics such as “adventurousness” or “achievement-striving.”
Once a sentry is surpassed, a warning is issued with a resolution option for the teacher. This resolution option allows educators to acknowledge that a sentry was breached and encourages educators to follow through with determining a solution for the issue. Sentry can help create and analyze data based on a learner’s written entries, which in turn, can be used to help educators predict a student’s mood or behavior. Sentry can also assist teachers in recognizing trends or signs of bullying or depression based on the moods of the students.
After a teacher is alerted with a warning, they can close it with a resolution note about how they went about either rewarding a positive trait, or providing a solution to a negative one. All previous warnings along with the notes are saved with the student’s profile history, which can provide educators with thorough evidence to reference at a later time, if any further intervention is necessary.
You can check out a limited demo version at sentry.mybluemix.net. The source code will be published at a later date.
—Andrew Lohr (@alohrAtIBM)