Time series discrepancies can provide insight into critical situations affecting a company's stock. However, detecting the cause of anomalies in a time series data is particularly challenging due to the lack of datasets and highly complex temporal correlations. The recent advancement of internet technology has transformed the communications network, allowing the public to disclose knowledge such as news, social media contents, and so on via the internet, resulting in the exponential growth of web data. Massive amounts of publicly available data may hold the key to unveiling the financial market's unexplained anomalous behaviors. In this paper, a review is done on anomaly detection in stock market series data and a study is done on correlations between sentiments of news, tweets and opinions and stock price which shows collective news sentiments inversely affect stock price. Further we propose an lexicon based approach improved on a financial dataset for finding influencing factors of an anomaly period. Our analysis shows a strong correlation between the sentiment of our data and the stock price and were also successful in extracting major events that could have been cause of on anomaly.