Do we generate insight by planning ahead rather than getting your hands dirty by jumping straight into data?
It all depends. On a very generalised view, it is best to dedicate the 'right' amount of time in each stage during an analysis. There isn't any right nor wrong, but there is only insane extremes.
I would argue that planning ahead is a waste of time if there is not even basic insight into the data. For instance, spending 1 day on planning an analysis (be careful that we are talking about a piece analysis, not the entire project) where where one does not even know what sort of type, quality and amount of data we are dealing with would just be a complete waste of time.
At another extreme, diving straight into raw data without purpose seems reckless. Yet we can often generate insights as we explore into data and hypothesis is usually drawn that way.
Hence, it does not harm to spend time 'faffing' with data in order to get familiarity with the data and to understand what we are dealing with. However, we have to know when to stop and pull back before browsing aimlessly and getting lost in the data exploration stage. The next wise step is to draw up a detailed plan for a thorough analysis.
To conclude, jumping straight in would not harm, but you have to realise at which point does it become marginally unproductive.
