It is all too easy for someone to see on a screen text written for a ticket and to handle it without emotion. However these tickets have been written by someone and that someone has emotional state.
What it does
Sentiment Analysis helps in two areas:
When using JIRA Service Desk, Sentiment Analysis can be used to understand the incoming tickets from customers and users. After a ticket is raised, the individual triaging it can clearly see the sentiment of the ticket and with a single click can see the breakdown of sentiment through the text the submitter wrote. This means that the triaging individual can prioritise the ticket easily and effectively based on both the priority and the sentiment. For instance a ticket using negative works from a customer such as "It is broken" or "Doesn't work" denotes a customer that is more unhappy and may need more rapid service than a ticket that says "Could we fix" or "It would be better if". Sentiment Analysis also helps if there are multiple people triaging tickets as it provides an initial standard base of understanding and reduces the chance that this information could be missed.
Improving Team Positivity
It is all too easy for a development team to write tickets and raise bugs with negative terminology. These negative comments are not good for team morale and constructive for the resolution of tickets. Sentiment analysis provides teams working in JIRA with the sentiment of what they have written, meaning that they can look to understand how they can write their tickets in a more positive way to boost moral and productivity. Again a team raising tickets saying "It stopped working" or "The login screen is broken" will experience more internal negative compared to a team where tickets are written such as "The login screen needs improvement". A simple exposure of a team to more positive tickets and sentiment will propagate through other parts of the teams day to day work such as stand ups, meetings and improve the team moral which in turn leads to happier work environment and more productivity.
It does this by analysing the ticket content including comments to work out the sentiment value and score of each ticket. It re analyses the tickets when they are loaded and will add to the score if comments and modifications are made to the tickets.
How I built it
I wrote this for atlassian forge in nodejs using VSCode and deployed it to forge easily and quickly.
Challenges I ran into
Working out how to deploy to forge and how the tooling worked.
What I learned
I found that Atlassian forge is a great way to extend the JIRA and JIRA Service desk. It provides a rich, infrastructure hassle free environment to build and extend. I will definitely be building more extensions.
What's next for Sentiment Analysis
Sentiment analysis would be extended to work for more languages with more advanced stemming and natural language analysis to understand phrases as well. This would make the application into an advanced sentiment analysis tool across Atlassian Cloud solutions.