We have been consulting as data scientists to an insurance company for the last 3 -4 months. We have taken their raw SQL outputs and programmed a lot of business intelligence in Python code.
We were done with our analysis about 4 weeks ago, however the client wanted to see the results in Excel so we were forced to jump through a few extra hoops and get our native python code to talk with Excel and output the magic. This all works fine – but!!
The client had to present our model to their Exco this week, and things didn’t go according to plan. For some reason they seem to do the odd manual data override on their system which comes out of their SQL report in an incorrect date format, which is screwing up our model’s output.
A good lesson in computer programming, you can be so near but so far.