About Mforecast

Mforecast is a Web-App written in R with the shiny package which allows you to forecast monthly time series based on the past observations multi-steps ahead. The technique includes statistical models and recurrent neural networks and is fully automatic. Any type of monthly equidistant time series can be handled by the algorithms. To make full use of the algorithms your time series must at least have 36 data points.

The design is such that you can get easy data up- and download. Zooming, renaming the axes and dynamic horizons are important features of this app.

Validation is an important task in every forecasting szenario. We provide an optical and data error view of the developement, what your forecast would have been, if you used this app! Compare it with your forecast generated in the past.

After forecasting you can download pdf (sample) and csv (sample) forecast reports quick and easy.

Please the Blog for instructions and further informations.

Moritz Müller-Navarra

Moritz Müller-Navarra is the founder and main developer of this app.

Moritz graduated in industrial engineering and management (equiv. Master of Engineering).

He wrote his thesis on forecasting sales with partial recurrent neural networks.

In his spare time he loves to play golf and football (soccer).

Please contact Moritz through moritz[at]mforecast.com or use google+.

Copyright © Moritz Müller-Navarra 2015