Blog | Reportbrain’s NLP for opinion mining: Going beyond sentiment analysis
Governments and Public Sectors nowadays need to be more and more prepared for a citizen-centric era. Improving the comfort of citizens’ life and making processes simpler and more automated has now become one of the main priorities. For this to happen, Local Governments need to always be aware of the number of matters people are facing and daily express their opinions about, so they could more easily address citizens’ needs for infrastructure and services changes in the city.
The biggest challenge comes on how it is possible to discover and analyse all this information, as it could be anywhere, from social media platforms to online reviews and conversations. Under the project of AI4PublicPolicy, Reportbrain has undertaken to create a tool that will support interactions between policy-makers and citizens, by providing data-based insights from citizen feedback analysis (using opinion mining and sentiment analysis). This technology aims to convert all the unstructured online information into a great asset in order for the policy-makers to be able to develop new, more responsive, and citizen-centric policies.
Technologies that we use
Citizen Feedback analysis is based on opinion mining. Opinion mining (also known as a deeper level of sentiment analysis and emotion AI) is the science of using text analysis (from structured, semi-structured, and unstructured data), to understand the drivers behind public sentiment/opinion. Opinion mining refers to the use of natural language processing, computational linguistics, and biometrics to systematically identify, extract, quantify, and study affective states and subjective information.
Reportbrain’s NLP for Opinion Mining can analyse citizen feedback from social media, online satisfaction surveys, chatbots among other sources, and provide those insights to policy-makers (local governments). Reportbrain’s technology is enabling organisations to see problematic areas to focus on by categorising the feedback into pre-defined customised categories. Also, it can visualise the sentiment around a citizen’s input, where sentiment will be described as positive/negative or neutral.
Additionally, Reportbrain’s language-specific models can recognize up to 28 emotions which can be detected on the sentence level. This feature is crucial to go beyond simple negative/positive classification and flag very negative comments/tweets/reports or other citizen input for organisations to react quickly and efficiently.
Reportbrain’s NLP for Opinion Mining is under development and Reportbrain is constantly improving its accuracy and precision. We trust that Citizen Feedback Analysis has a very high potential and will turn out to be of great support to the Public Organisations and Policy-makers. In this ever-changing world of information and communications, Local Governments need to acknowledge the importance of taking advantage of new AI technologies, in order to enhance citizens’ experience when it comes to public services. And by using Reportbrain’s NLP for Opinion Mining they get closer to achieving that goal.