0
$\begingroup$

Pardon, I am a novice of time series analysis.

I have merged together meteorological data from two different sources, one derived from the Brazilian National Institute of Meteorology (INMET) and the other derived from the OpenWeatherMap API (https://openweathermap.org).

The INMET data was drawn from a selected city which provides current hourly readings for a number of weather parameters (e.g. temperature, humidity, rainfall and pressure). The OpenWeatherMap data was also drawn for the same city through an API. The only difference for the OpenWeatherMap is that it provides current 3-hourly readings for the weather parameters which are also forecasted ahead of time in a 3-hourly interval up to 5 days.

I compiled the data from 11-March-2020 to 25-July-2020 and want to compare the predictions derived by the OpenWeatherMap against the actual values observed by INMET to examine whether the forecasts from OpenWeatherMap closely match those from INMET. The structure of dataset are of time series. However, I am not sure what is the best approach (i.e. MAPE, ARIMA, ARMA and dynamic time warping).

I have attached a link to an example dataset to illustrate what the data looks like which contains actuals and predicted (i.e. future 3hr, future 6hr, future 9hr and future 12hr) readings for temperature from INMET and OpenWeatherMap, respectively.

Dataset: https://drive.google.com/file/d/1zuQ9ympJV0H5Nf75tHLBjNogadhbo1Ip/view?usp=sharing

Please any help, advice or instructions for performing this type of analysis would be appreciated.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.