If you've gone through all the recipes in this chapter so far, you might need a break from atmospheric pressure. So let's look into visibility instead. The data file has a column for a maximum visibility, which the KNMI describes as follows:
"Maximum visibility; 0: <100 m, 1:100-200 m, 2:200-300 m,..., 49:4900-5000 m, 50:5-6 km, 56:6-7 km, 57:7-8 km,..., 79:29-30 km, 80:30-35 km, 81:35-40 km,..., 89: >70 km)"
Visibility here is a discrete variable, so averaging values may not make sense. Also, it seems that we have a lot of 0 values for the period between 1901 and 1950 for almost every day. I don't believe that De Bilt was extra foggy in that period. For the purpose of this recipe, we define mist as visibility between 1 and 2 km, which corresponds to the values of 10 and 20 in the data file. Let's also define haze as visibility between 2 and 5 km. This in turn corresponds to 20 and 50 in our data file.
Air pollution could reduce visibility, especially on clear...