Predict for how long the money available right now can support the user/family

[htmfilho]

Based on the history of transactions, which represents the spending habits of a user/family, predict for how long the money available can support the consumption pattern. The output should be the number of days left until running out of money.

The algorithm should ignore user's income transactions in all predictions, so it always considers the worse case scenario, which is when all incomes stop suddenly.

It is also important to consider seasonality because we spend more in some months and less in others. For example: we may spend more in August with school supplies and less in September when the routine gets back to normal. More in December because of Christmas and less in January. Etc.