**2017 2018 2019 2020**
16 31 147 123
23 33 102 116
13 10 128 86
32 27 128 70
48 70 164 95
26 41 113 131
34 26 179 82
23 38 197 171
37 41 156 320
41 66 160 115
31 63 117 106
27 48 186 103
This is data collected between Jan - Dec of 2017 and 2020. As can be seen, more effort went into data collection in 2019 and 2020 while little was done in 2017 and 2018. Will it be statistically sound to normalise each year separately and merge all four years into one column (for uniformity) as a feature in a dataframe?
Each year is a representation of collection. I am worried about the impact of 2017 and 2018 on my model performance. My intention is to predict collection for 2021 or 2022 etc.
