The presentation will explore the potential of integrating natural language processing (NLP) techniques with time series modeling to improve predictions of transnational financial flows, taking remittance inflows to India as a case study. My research employs a mixed methods approach combining quantitative SARIMA time series modeling with qualitative sentiment analysis of news articles related to the Indian diaspora and remittances. An initial SARIMA(1,2,1)(1,1,1,12) model yielded a high mean absolute percentage error (MAPE) of 62.96% for remittance forecasts. To enhance accuracy, sentiment scores derived from NLP analysis of 1000 news articles were incorporated as adjustment factors into the SARIMA model. This integration dramatically improved the MAPE to 6.25%, highlighting the benefits of overlaying NLP-based qualitative factors onto statistical forecasting models. The research demonstrates the value of an inter- disciplinary approach combining computational linguistics and econometrics to create more robust predictions of complex financial flows. It opens up new possibilities for integrating textual data analytics into financial forecasting.