Use case:
How likely stock A will [go up] TODAY, given all stock data for YESTERDAY

Make sure that all ML processes are working correctly

Sz should over current python code so the team is on the same page
Team need to work on system design for ML

Use case:
Pick a stock A. If A [went up] YESTERDAY, find all the stocks that will [go down] TODAY
Requirement:
Prediction much be greater than 75%
For relation that < 75%, label as unrelated

http://scikit-learn.org/stable/modules/model_persistence.html

Use the diff between Open and High as the features to predict the next day price of a stock
Can add volume as weight

create redux store for UI

@lq3297401

maybe change day picker font, because the circle background now is not a circle

Code free
Debug and make sure that our programs are bug-free

How should we build the ML system?

Same as Normalized Data Python. But result should be store in DB. On table update, the normalized data should be calculated automatically.

Currently, we have data in DB in the form:
{ Date, Open, High, Low, Close, AdjClose, Volume}
We should normalize the data before doing ML.
A python function to return normalized data is needed.
Normalized data should look like this
{ Date, (Open(Yesterday) - Open(Today))/Open(Today) ....)

add comment to your python code so it will be easier to understand for other teammate to learn from your code
http://sphinxcontrib-napoleon.readthedocs.io/en/latest/example_google.html

Rename files so they're easier to understand
Factor out reusable codes
Setup framework to make future improvements easier
Add some comment to help others understand your code

need prediction history data
we will show the prediction history data in the line chart in the prediction page

Block by Normalized Data.

After data is normalized, we should be able to 'group' stocks together.
We should only care about Close Price to make it easier. Stocks that are always Close '+' should group together, and Close '-' group together.
I still don't know how many group we will have and how to do this. We may use this http://scikit-learn.org/stable/modules/generated/sklearn.cluster.MeanShift.html#sklearn.cluster.MeanShift

http://scikit-learn.org/stable/modules/svm.html#regression

Use SVMRegression to predict the next date price of a stock
1-Use single stock to predict it next date price
2-Use multiple stocks to predict a stock next date price

bugs in date picker "date from", the triangle for previous month not working

Use the diff between Open and Close as the features to predict the next day price of a stock

Use the diff between Open and Low as the features to predict the next day price of a stock
Can add volume as weight

go over the code in server folder and work with Sz to add more comments for future development
https://milmazz.uno/article/2014/08/27/how-to-document-your-javascript-code/

Show current stock name on the UI

http://scikit-learn.org/stable/modules/sgd.html#regression

Use SGDRegressor to predict the next date price of a stock
1-Use single stock to predict it next date price
2-Use multiple stocks to predict a stock next date price

move to mysql-cache
https://www.npmjs.com/package/mysql-cache

Go back to your code and add more comment
https://milmazz.uno/article/2014/08/27/how-to-document-your-javascript-code/

Learn how to setup, dev, run Python project

Make the current date time picker works

Use case:
Pick a stock A. If A [went up] YESTERDAY, find all the stocks that will [go up] TODAY
Requirement:
Prediction much be greater than 75%
For relation that < 75%, label as unrelated

Use case:
Given the history of a stock A in the pass up to YESTERDAY, how likely A will [go up] TODAY
Requirement:
Prediction much be > 70%

Create an API to get all stock symbols
Basically, just query the 4update table and get the Symbol column.

New requirement for server: accept get request in form
Base: /Stocks
table=____ String: table name
startDate=____ String: utc time
endDate=____ String utc time
http://54.219.174.5:9081/Stocks/AAL/2012-01-02/2013-01-02
become
http://54.219.174.5:9081/Stocks?table=AAL&startDate=2012-01-02&endDate=2013-01-02
Error handling:
If table is missing return empty
If startDate is missing, set startDate = epoch 0 = 1970-1-1
If endData is missing, set endDate = now()
Additional feature:
Server should accept startDate or endDate in epoch time second form

UI should have an option for a user to change the stock symbol to be displayed on the current UI
For prototype, a simple text box with a submit button is good enough

Use the diff between Open and Adj Close as the features to predict the next day price of a stock
Can add volume as weight

Code free
Fix or improve code quality