Locational marginal prices of electricity and weather conditions in Yucatan peninsula

Published: 12-05-2020| Version 1 | DOI: 10.17632/7ckhh8hc2h.1
Adrian Livas-Garcia,
Emmanuel de la Cruz May,


It is presented two datasets used to train a neural network that forecasts electricity prices in the Yucatan peninsula. The first one is the Input data, which is composed of five parameters, three describing environmental conditions and two reporting the levels of operation of the electricity system in the study region. The second is the output data, corresponding to local marginal electricity prices. These prices are compound from the next three costs: energy, losses of transmission, and congestion. Also, these data allow detecting the dynamics of the electricity market, which can be related to environmental conditions. Also, they allow detecting phenomena of the electricity market, i.e. negative prices of transmission losses or congestion, and the negative merit-order effect. Every parameter was collected for eight load zones in hourly resolution, it is the geographic distribution according to the Mexican independent system operator. The data begins in the first hour of January 1st of 2017 and ends in the last hour of April 4th of 2019. Each parameter has 157808 observations.