# Most effective way to learn time series with poor quant background

End goal will be practical application (model building) by using time series analysis to analyze/forecast macroeconomic/finance data.

Background: I have taken stats, introductory econometrics, intermediate macro, micro in the past couple years, but is has been 10 years since I took my only calculus course.

I'm confused whether I should (a) go do all the required math, (b) just find a math econ refresher course, then to read a book focused on theory or (c) just try to build models.

I realize I'm starting from a low base, but I'm hoping to become reasonably proficient in a year (unless this is unrealistic?).

What, in my opinion, you need to do is first study difference equations, then basic statistics and after that you can approach the elementary time series models, e.g. ARMA, GARCH and so on. Math and Stats are a prerequisites I am afraid.

A good math book for econ students is "Fundamental Methods of Mathematical Economics" by Chiang, Wainright. I am sure there are others but that one I use and know very well. You will get the knowledge you need about difference equations there.

First build models.
(1) Decide on the statistical package you're interested in using.
(2) Choose a standard practical textbook using this statistical package. (Stata: http://www.amazon.com/Introduction-Time-Series-using-Stata/dp/1597181323/ref=sr_1_1?ie=UTF8&qid=1386268375&sr=8-1&keywords=stata+time+series)

Then deepen your understanding of methods based on time available to you.
(3) Based on interest, and uncertainties from reading textbook, delve deeper into specific areas using primary/secondary literature.

You have sufficient math background, in my opinion. I would suggest Shumway & Stoffer's book on time series. It comes with R package, complete with code examples and data. There is not a lot of math involved. I'm sure you'll enjoy the book. It is most important to start working with statistical packages and data. Once you build intuition, you can later on dive into the depth of statistics.

First, don't use people terminology like "quant".

Second, do a bachelor's degree in mathematics or stop thinking you will ever do model building or anything like that.

Mathematics is the foundation of almost all technical and related fields. You need it; otherwise work with a mathematician.

• Mathematics can definitely be useful but it's certainly not true that everyone working in applied statistics, economics, modeling or related technical fields has either a degree in pure math or access to a card-carrying mathematician. So what's the claim here? – Gala Jun 10 '14 at 9:18