We identify the significant independent variables for mid-term electricity demand forecasting with hourly intervals, in the Netherlands. We show that meteorological, historical and calendar data should be used for mid-term forecast. The independent variables Dutch Carnival, school and national holidays and Bouwvak have different effects on the electricity infeed in different regions. The effect of a holiday also significantly depends on the day of the week and the hour of the day. For instance, a holiday on a Sunday only has an effect on electricity demand during the day, while the effect of a holiday on a weekday or Saturday is significant for all hours of the day. Finally, we point out that, holidays also affect the electricity infeed on the day before the holiday and the day after the holiday. However, this effect is also conditioned by the hour of the day, and hence each hour of a day should be treated separately in the forecasting models for DSOs.