This is the most important part of any data analytics project. Once you have your data cleaned and properly prepared to feed a training algorithm, you have just to choose which machine learning or statistics based algorithm to use.
Below is a list of the most commonly used algorithms, old and new, supervised and unsupervised, coming from the statistics or from the machine learning community, predicting numerical values or nominal classes, requiring past time trends or just past examples.
After viewing the videos and learning how they work and what they can do for you, choose the one that best fits your dataset and your business problem. If you are not sure you can always train a few and make them compete for best performances or you can make them work together in an ensemble model.
As you see, there are many options.