Quantitative Financial Data analysis

• Please see the attached csv file for the sample dataset.
• The goal of the project is to examine a time series, i.e. the “Signal” column, whether it has
predictive power of future returns of SPY, i.e. “ClosePrice” column.
• The “Signal” is generated at the end of the day, the same time as SPY’s close arrival. We
don’t have prior assumption on prediction horizon over which the signal is effective.
• The dataset may contain data errors, e.g. missing data, outliers, date error etc. Please feel
free to apply reasonable assumptions to clean the data if it helps your analysis.
• Please submit code, notes, graphs, or any documents that conclude your research (i.e. does
the signal predict future SPY return?) and please explain your analysis.

PROJECT 2: Portfolio Optimization
• Given a universe of 500 stocks
• Their alpha forecasts are a_i {i=1, 2, 3, . . . , 500}
• The covariance matrix of the stock returns is V, where v_{i, j=1, 2, 3, . . . , 500} represents covariance
between stock_i and stock_j
• For trading cost, let only consider market impact, and assume it is proportional to the trade
size and a constant coefficient applying to all stocks.
Please make up the inputs for stocks alphas, covariance, and any needed parameters,
e.g. risk aversion, portfolio size, etc. (please briefly describe your choice of the settings).
Formulate a single-period mean-variance optimization problem with trading cost.
Either long/short portfolio or long only portfolio is fine. Solve it numerically, and submit
code, graphs, and notes that help explain your work.