During disasters, it is extremely crucial that the right resources are received to the victims within time. Disaster relief NGO's revealed that there is often mismanagement and lack of coordination. Also, often identifying right resources takes up a lot of crucial time. To aid this, we developed an algorithm that identifies locations from microblogs, upto 100x faster than SoTA StanfordNLP.
We first developed an algorithm to identify location from microblogs in a real time situation, which was 100 times faster than state of the art StanfordNLP. The proposed algorithm is also faster than other tools. We used NLP tools like dependency paring, Named Entity Recognition, and other rules to identify the location. This resulted into a research paper at WWW2018, WebConf held at France. To further assist disaster relief attempts, we developed a platform that could identify crucial information like Resources, location, quantity etc thus effectively using social media to aid disaster mitigation. The proposed platform emerged first in Microsoft's code.fun.do Hackathon out of 242 participants, and was one of 21 student projects to be demonstrated on a national level at Microsoft's AXLE, 2019.