As a country who had a high risk affected by the earthquake, social media have an important role. Besides to serving earthquake information, the spread of information on social media is so wide and fast. However, information on social media has a gap to reach validity and doesn't contain detailed information about spatial information. By leveraging crawling result data on Twitter, then data will be processed with Natural Language Processing (NLP), this research aims to proves about transformation of unstructured data into structured data with NLP for use on spatial analysis in Indonesia using data text on platform social media, Twitter. In addition, this research is also aims to reveal correlation between earthquake magnitude and earthquake frequency. The results proves that NLP can be used for spatial analysis with data text on Twitter related to earthquake. Besides that, the value of maximum magnitude are great significance to the earthquake frequency.