Solar Power Generation Data

https://www.kaggle.com/anikannal/solar-power-generation-data
There are a few areas of concern at the solar power plant:

Exploration of this dataset with several weird data values, and other oddities, that lead to finding some issues with the hardware of a plant. Further digging leads us to a conclusion: plant X has systemic problems.

Features:

I guessed that the values of total yield for plant 2 were not comparable with those of plant 1 in the original form. So I took a guess and used a unit conversion formula: 1kWh = 3.6⋅106J = 3600kJ = 3.6MJ = 0.0036GJ

solar1GEN['DATE'] = pd.to_datetime(solar1GEN['DATE_TIME'],format ='%Y-%m-%d')

We drop the PLANT_ID's to spare memory, and rename the inverter labels.

First rename these strange labels

Limit temperature and irradiation values' level of precision