Dr. Shepherd's research focuses on software engineering, human computer interaction, and computer science education.  His applied research has consistently resulted in working prototypes and products which serve to provide real world validation of academic ideas.  Current research topics include: end-user programming for industrial machines and increasing diversity in computer science. ​

current projects:

End-User Programming of Industrial Machines
As industrial machines, such as collaborative robots and PLCs, drop in price, they become affordable for small business and "makers".  Unfortunately, due to their dated user interfaces and arcane programming languages, they are nearly impossible for non-experts to program.  Our work focuses on creating modern interfaces that enable novices to program industrial machines, thereby making these machines accessible to a wider audience.  
Increasing Diversity in Computer Science
Though there have been many efforts to reverse the trend, computer science continues to suffer from a lack of diversity.  Our work creates a culturally-relevant curriculum---using code to create hip hop beats---in an effort to increase engagement of minoritized youth in computer science during the crucial middle school years.  Sound interesting?  We will be running a camp this summer!  Contact Dr. Shepherd for more information.

completed projects:

FlowLight
Interruptions devastate the productivity of office workers. To combat interruptions, we created the FlowLight, a physical traffic-light like LED combined with an automatic interruptibility measure based on computer interaction data. In a large-scale and long-term field study with 449 participants from 12 different countries, we found, amongst other results, that the FlowLight reduced interruptions of participants by 46%, increased their awareness on the potential disruptiveness of interruptions, and most participants are still using it today!  The FlowLight was featured in the Wall Street Journal and is now available for purchase.
Code Search
Code search stinks.  There's no auto-correct or suggestions, regex-based searches fail most of the time, searching for two terms is nearly impossible, and the returned results are unranked.  Sando is built on top of Lucene so it provides ranked results, multi-term search, and near instant results.  It leverages natural language processing to provide code-appropriate auto-complete and uses software-specific synonyms to provide suggestions.  Sando has been downloaded by over 48,000 Visual Studio users and has enjoyed widespread usage.
​