I made a screencast about how to read a CSV using pandas. There are always things that go wrong when reading in data (date parsing, missing data, bad encodings…), so I went through fixing a few of them.
Final result: http://nbviewer.ipython.org/484ad73a7208066d6aec
Data I used: http://depot.ville.montreal.qc.ca/velos-comptage/data.zip