The idea behind this poster is to show to the viewers that we can/must use Python (loaded with some scientific “batteries”) to do quantitative research in the financial “arena”. To achieve this objective, we propose us to explore the integration of specific tools like Numpy, Scipy, IPython, Matplotlib, Statsmodels and Pandas, to do time series analysis and modeling with "real" financial time series, in particular, we will show you how we can develop a statistical arbitrage strategy - pairs trading - implementing the search of cointegrated financial time series, characterizing and modeling the resulting stationary “spread” and evaluating the performance of our strategy in historical data (back-testing) and in the “real" market (forward-testing). Our final aim is to show to the viewers that we can use Python like a quasi-complete environment to do this kind of research, enlightening the pros and discussing the cons, and what we can hope in the future in this area.