%% Lecture 2.
%
% READING DUE NEXT WEEK:  Magrab, skim Chapter 1, read 2.1--2.5
%
%  HW: Magrab, 
%     problem Section 2.3, exercise 2.3 (a,b) on page 79, and 
%     problem Section 2.5, exercise 2.12 on page 81
%   due at beginning of class next week.
%   turn in a print of of a script file with the commands used to solve
%   these problems.
%     time: 5-30 minutes.
% in class next week, a closed book and notes quiz will ask for hand
% written answers to 2 questions on creating vectors and matrices.
%
% LAST WEEK:  I tried to demonstrate that matlab is a very powerful tool
% that does math operations very well... with a focus on numerical
% methods rather than symbolic computation (which is more the domain of
% maple and mathematica).
%
% OPENING QUIZ:
%   1. write (on paper) one matlab command that would make a list of
%         integers from 1 through 50
%   2. write (on paper) one matlab command that would 114 evenly spaced
%         numbers starting at 14 and ending at 22
%   3. write (on paper) one matlab command that would make a list that
%         looks like:   [100 99 98 97 96 ... 1]
%
% TODAY: My goal is to completely work out a few specific examples
%        I hope that you enjoyed the Mandelbrot Set Laboratory.  Like the
%        first lab, that had many steps that we went through
%        together... and some steps (some lines that you typed in) did
%        not have an obvious purpose until later in the lab.  In this
%        lecture we will try to focus a little bit more on how to
%        organize our approach to a problem, as we teach a little bit
%        more syntax of matlab.
%
%  So let's start today with a problem, and since so much focus in
%  technology development is based on the profits and returns, lets first
%  focus on money.  Let's do an analysis of recent stock prices --- in
%  particular, lets find days when both google and yahoo's stock prices
%  went up.
%
%  We could jump in here and start, but it is useful to have a strategy
%  in mind for how to proceed, and it is good to have a meta-strategy for
%  how to create a good strategy.

% Meta Strategy:
%    1. Specify the problem exactly.
%    2. Define and do any programming necessary to get the input data.
%    3. Then define the strategy for how to address the problem
%
%  (why this?  you should not work too hard to plan your approach to the
%  problem until you know what data you can get).
%
% Meta Planning:
%
%    1. the goal is "lets find days when both google and yahoo's stock prices
%       went up".  More precisely, "Given the stock price data every day,
%       find which days the price of both yahoo and google went up".
%    2. The input data:
%       So much is available online these days.  We can get the data we
%       need from:
%           http://finance.yahoo.com/q/hp?s=AAPL
%           which has a link at the bottom to save the data to a
%              csv spreadsheet (csv == comma seperated values)
%           This is a very common, portable file format, and many
%           languages already have functions defined to read them.
%       In matlab, the function is:

google = csvread('google.csv',1,1);
yahoo = csvread('yahoo.csv',1,1);
% important that this is in your "path" --- the current "relevant directory"

% the 2,2 at the end says that we want to skip the first row and column,
% so that we get just the matrix of numerical data.  type "help csvread"
% to get more information.

% The "closing value" of the stock is in the fourth column, so lets
% grab just that column and put it into a new variable:  
googleClose = google(:,4);

% and do the same for yahoo.
yahooClose = yahoo(:,4);

% now, let's plot these 
plot(googleClose);
hold on;
plot(yahooClose);

% did google really go down in this time?
%... This array is arranged backwards in time.It'd be nice to fix that:
googleFixed = googleClose(end:-1:1);
yahooFixed = yahooClose(end:-1:1);

% So, now we have our data values, 
% lets make a strategy for finding when they both rise. 
%   1. Convert the Google data into positive and negative percent changes.
%   2. Convert the Yahoo data into positive and negative percent changes.
%   3. Find out when Google rose.
%   4. Find out when Yahoo rose.
%   5. Find out when they both rose.

%  Now we can implement these step by step.
% we should also check that we get each step right before we go on. (!)
% step 1?
%   1. Convert the Google data into positive and negative percent changes.

%  percent change (in general, not in matlab terms) is:
googleToday = googleFixed(2:end);
googeeYesterday = googleFixed(1:end-1);
googleChange = (today-yesterday)/yesterday

googleChange = percentChange(googleFixed);
% what will this look like for yahoo?

% if they are so similar, good excuse to write a function!
 % why is writing a function good?
 
 %matlab function definition, in file "NAME.m"
 % function answer = NAME(arguments,...)
 % do stuff...
 % do stuff...
 % answer = something...
 
 % when that function is written, we can call it with:
 yahooChange = percentChange(yahooFixed);
 googleChange = percentChange(googleFixed);

 
 % Now, how can we tell when google is rising?
 
 %suggestion...
 googleRises = googleChange > 0; % googleChange(1:end)>0, googleChange(:) > 0
 
 % and for yahoo?  If it is similar, another chance for a function!
 
 yahooRises = yahooChange > 0;
 
 % then lets find when they both rise.
 
rises = (yahooRises & googleRises) % gets printed
days = find(rises) % gets printed; indexes of days.

 %%%% This is the end of what was asked for.  But... now that we are
 %%%% here and have all this nice data...

% How could we find days where both stocks fell?


% how can we do this in just one line?


% How can we display the correlations better?


% tools for labelling plots
xlabel
ylabel

% New commands today:
% csvread
% &, |, ~
% find