{"This is an hourly electricity consumption data with 8784 records for the whole 2016 year. The following are the attributes in the energy_consumption_levels csv file. 3_levels: three levels (i.e. low, med, high) of energy use 5_levels: five levels (i.e. very low, low, med, high, very high) of energy use 7_levels: seven levels (i.e. level 1, 2, 3, 4, 5, 6, 7) of energy use in MWh consumption: numerical value of energy use temperature: historical hourly meteorological data (temperature) measured in Celsius degrees hour_of_day: hour of the day (1 to 24) day_of_week: day of the week (1 to 7) day_of_month: day of the month (1 to 31) month_of_year: month of the year (1 to 12) The energy consumption is retrieved from smart metering device from a commercial-type consumer and the environmental temperature is recorded by the meteorological sensors of a specialized institute for the whole year 2016. The energy levels (3_levels, 5_levels and 7_levels) are produced by cutting the data into n number of bins (3, 5, 7) using simple statistics or using some already prepared functions. The rest of the data are already cleaned originally prepared by Pîrjan, Alexandru, et al. in their paper entitled: \"Devising Hourly Forecasting Solutions Regarding Electricity Consumption in the Case of Commercial Center Type Consumers.\" (published in Energies 10.11 (2017): 1727) with the new added attributes on its energy levels."}
