Traffic congestion is a huge problem in urban environments causing immense losses of time and money. Using time series data collected over a month, this project conducts an exploratory analysis for identification of patterns in traffic movement. The data is a combination of two New York City sources - CCTV cameras images and local weather metrics. Using the results from the analysis of this combined data, a neural network model has been built that can predict traffic states.This model may potentially lead to more dynamic traffic signal change systems.