ses_portfolio

This project exists to exercise elasticsearch with scala elasticsearch client, elastic4s. The examples in this project are inspired by problem sets in the Coursera course, Computational Investing, Part I.

In particular, current examples use elaticsearch to :

• Create portfolios with N equities and benchmark indexes as elasticsearch indexes using historical price data from Yahoo Finance, downloaded as CSV files. The lines in the CSV files are translated into JSON using CSVParboiledParser by Juergen Pfundt. The portfolios created are:

  1. Simple elasticseatch index containing the following document with daily prices for each equity, e.g.:

     { "name": "msft,     
        "date": "2011-01-3",
        "open": 30.25,
        "high": 30.52,
        "low" : 30.15,
        "close" : 30.28,
        "volume": 1234567,
        "adj_close: 30.26 
      }
     

  This index is used to calculate sharpe ratio examples as explained in the following bullets.

  2. A simple elastic search index whose documents contain the daily rate of return (ex dividend) for four equities,e.g.:

     { "date" : "2011-01-04",
        "msft_ror" : 0.001345,
        "amzn_ror" : 0.004589,
        "ebay_ror" : 0.0145,
        "ups_ror"  : 0.00111
      }
     

  This index is used to maximize the sharpe ratio of a four equity portfolio.

  3. A simple elastic search index whose documents contain the weekly rate of return (ex dividend) for Micsrosoft (MSFT) and S&P 500 index equities,e.g.:

     {  "date" : "2011-01-14",
        "msft_ror" : 0.001745,
        "snp_ror"  : 0.003547
      }
     

  This index is used to calclulate for MSFT the beta coefficient of the Capital Asset Pricing Model (CAPM).

4. An portfolio positions index that records all a portfolio's asset positions val for each trading day whithin one year,e.g:

 {
    "date" : "2011-01-14",
    "cash" : 987876.00
    "aapl" : 123454.00
 }
 ```
This index is the result of a trading simulation and used to calculate the portfolio's: average daily rate of return, cumulative return, standard deviation, and sharpe ratio

* Calculate the [sharpe ratio](http://en.wikipedia.org/wiki/Sharpe_ratio) for individual equities using the elasticsearch [extended stats aggregation](http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/search-aggregations-metrics-extendedstats-aggregation.html);
* Calculate the sharpe ratio for multiple equities using elasticsearch [terms aggregation](http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/search-aggregations-bucket-terms-aggregation.html) to create one bucket per equity and then using extended stats aggregation in each bucket to calculate the sharpe ratio for each equity.
* Calculate the weights of allocation for a four equity portfolio that maximize the sharp ratio for the portfolio as a whole. This example uses the elasticsearch [script field](http://www.elasticsearch.org/guide/en/elasticsearch/reference/current/search-request-script-fields.html), to distribute the allocation weights amongst the closing prices. 
* Calculate [beta CAPM coefficient](http://en.wikipedia.org/wiki/Capital_asset_pricing_model)for MSFT against the S&P 500 over a 3 year peropd from 2011 to 2014. The calculation uses a linear regression algorithm implemented in [scalanlp](http://www.scalanlp.org), in particular, with the [breeze](https://github.com/scalanlp/breeze) and [nak](https://github.com/scalanlp/nak) components.
* From a set of trading orders and equity adjusted closing prices, calculate a portfolio's average daily rate of return, cumulative return, standard deviation, and sharpe ratio.

### How to Run Examples

From the commandline run the following sbt command: 

 sbt "project analytics" "run <args>"     

The initial response from sbt will be:

_Multiple main classes detected, select one to run:_

_[1] org.hfgiii.sesportfolio.analytics.AnalyticsMain_

_[2] org.hfgiii.sesportfolio.analytics.client.esexamples.EsExamples_

_Enter number:_

If one enters _2_, there is no need for arguments and `<args>` is empty. The first selection, _org.hfgiii.sesportfolio.analytics.AnalyticsMain_, requires arguments. The following table shows the valid input:

| Option   | ValidValues | Calculation |
| :------: | :---------: | :---------: |
|   o      | true,false  | optimize portfolio |
|   s      | true,false  | calc sharpe ratio |
|  e       | msft,amzn,ebay,ups| combine with _s_ for single sharpe ratio calc|
|  b       | true        | calc CAPM beta for MSFT|
|  t       | true        | trade simulation|

Notice: If one enters the _s_ option without an _e_ option, the sharpe ratio is calculated for four equities: _msft, amzn, ebay, ups_. The _e_ option by itself with a valid equity is a __NOOP__; the example will exit silently. If the equity is not valid, it responds with following message:" '__invalid equity name__' is not a valid equity for sharpe ratio calculation"

Issue: when sbt runs the examples, it will throw a stack trace most of the times when the example code shutdowns elasticsearch. Will fix this shortly.