is creating R Return Analysis (FRED & Stocks)

Select a membership level

Support Tier
per month
Access to support on utilizing opensource code and setup for presentation server.
Report: Monthly Returns
per month
Publish a pivot list of monthly returns sorted from highest stock to lowest
Report: Weekly Returns
per month
Publish a pivot list of weekly returns sorted from highest stock to lowest.

Publish a fact table of monthly adjusted closing prices per stock




per month


Using principles learned from Data Science Masters program (Kimball/Data Warehouse Lifecycle Method) my goal is to develop a Data Warehouse system from ETL to Presentation server, eventually putting it all online. Providing static reporting services initially, to eventually expanding to aggregate reports and underlying services that support ad hoc querying (cached queries over aggregated metadata). The presentation server will hold financial data extracted from St. Louis Federal Reserve as well as Stock Market data using publicly available API's and pulling the information into R for advanced analysis.
Disclaimer: The presentation server will process data inputs extracted from FRED and other API's, but will not hold data directly from the FRED or API's, as this is a violation of some API's (from what I gather from reading the API use agreements, use of data is okay as long as it's transformed from original and the original is not shared).  Any and all such data will be cleaned and transformed on input.

Eventually a presentation server will allow for third party tool access. Initially through exported reports via user requests. But soon to be direct access via hosted presentation server ports.

Then build further analytical capability (such as advanced R statistical analysis on the ingested data sets**) into the system in terms of feature requests. Feature requests will be derived from the community via forums/approval voting via a community maintained feature request list with estimated value and feasibility of each request which constitutes the prioritization grid (an x,y matrix) (prioritization grid, see example:

The goal is to maintain a steady flow of updates in terms of commit changes and feature releases. The plan is to be working on 3-5 features at any given moment releasing 1-3 features a month. This is entirely dependent upon the feasibility (i.e. level of effort of the features requested).

Announcements will be made via Patreon page of qualified features. The goal is to only work on features within the upper right quadrant of a prioritization grid. Of those qualified features, further discussion will follow on the community forum. Followed by a vote. This forum will also serve as community support for development as well as implementation.

Presentation server code will be open source.  I'm not sure how I intend on keeping my code changes private, but it certainly won't be on github.  I think I'll have to switch to either a local repo or a different private repo.  If this were just the community I'd be worried about, I'd share the code, but I'm more concerned about competitors monetizing off of my hard work by replicating my code.  Patreon isn't so much about opensource, but about skilled aristans sharing their work.

So with that said.  What I am willing to share is code to a presentation server base.  Beyond that I hope to offer access to reporting services, such as pivoted tables of daily adjusted returns, weekly adjusted returns, monthly adjusted returns per stock sorted by average return with rules to detect outlier's and filter inpure data sets (more than 1% missing values).

Beyond that maybe basic portfolio services.  I think most people would simply like access to the return matrix.  I think to incentive the work, I could share an aggregate, either useful or not. Such as yearly vs weekly returns.  That would truncate the size of the shared data considerably and create repoire as well as demand for more.

When it comes to choosing among qualified features, the plan is to give higher tier members more weight in terms of rank as incentive to contribute more.  Maybe even set up a two split system between what money says and what the individual people say and weight them accordingly.  In the end all things considered even, money talks 

The intent is to use feature requests as the selling point, to offer higher tier members better access to features they may desire. I will most likely setup an approval voting method where higher tier members get vote values proportional to their monthly contribution amounts. This model will evolve based upon community sentiment.

Code on github
Current Features
* Multiple Regression
* Binary Logistic Regressoin
* Sliding Windows
* Ability to apply queries and derive returns

Users Vote via:

Final Feature Report


Terms of service of intended API's:

Alpha Advantage
Note: "Raw", to work with this requirement. The ETL will parse "RAW" entries at the source/ETL process and send record by record vs sending any raw CSV's in full.
By becoming a patron, you'll instantly unlock access to 1 exclusive post
By becoming a patron, you'll instantly unlock access to 1 exclusive post

Recent posts by ReadyLoop

How it works

Get started in 2 minutes

Choose a membership
Sign up
Add a payment method
Get benefits