Author : Miss. Ashwini Gaikwad and Prof. Pravin Nimbalkar

Poorly maintained roads are a fact of life in most developing countries including our India. A well-maintained road network is a must for the well-being and the development of any country.  Our work main focuses to create an effective road surface monitoring system with help of one of social media platform. Twitter is a social networking service with more than three hundred million users, producing a large amount of facts each day. Twitter’s most important characteristic is its ability for users to tweet(message) these tweets are all about events, situations, feelings, opinions, or even something totally new, in real time. The social media tweet has textual content which are mined that allows to perceive the lawsuits regarding various avenue transportation issues of site visitors, such as accidents, and potholes. In order to pick out and segregate tweets associated with exceptional troubles, keyword-based methods have been used formerly, these strategies are entirely depending on seed key phrases which can be manually given and these sets of keywords aren’t sufficient to crawl through all tweet’s posts. So, to conquer this issue, a singular approach has been proposed that captures the semantic context via dense word embedding by employing the word2vec model. However, the system of tweet segregation on the idea of semantic comparable key phrases may additionally suffer from the problem of pragmatic ambiguity. To take care of this, Word2Vec model has been implemented to shape the semantically similar tweets. Furthermore, the hotspots had been identified similar to each category. As there is scarcity of geo tagged tweets, we have proposed a hybrid method which amalgamates Named Entity Recognition (NER), Part of speech (POS), and Regular Expression (RE) to extract the vicinity statistics from the tweet text. This work will assist to avoid injuries and can use to become aware of trouble areas ahead of happening. The government may be alerted to take preventive movements and these preventive actions can save money. Our work contributes to automation of pothole detection using Android mobile, Google map, microcontroller, GSM modem, GPS, Ultrasonic sensor making it liable to be used for social welfare.

Keywords: Travel Habits, Road Conditions Detection, Road Traffic detection, Incident Detection,  Social Media, Named Entity Recognition, Word Embedding, Transportation, Pothole Detection.