1 of 26

Stock Market Shiny app

Yicheng Shen, Zheng Ren, Angela Zhao, Yuzi Li

2024.12.12

2 of 26

Requirement

Presentation quality: Is the presentation clear on what was the question being investigated? What were the original goals and what was accomplished along the way? Was there previous work in this area? Are the slides readable? Do the figures have large enough legends and figure titles? Did you describe the data?

Paradigm integration: How effectively does your project demonstrate multiple paradigms?

Functionality: Does the data analytic product work as intended? Are the components well-integrated?

Usability and documentation: Is the data analytic product easy to use/read and well-documented?

Originality and complexity: Does the project address a non-trivial statistical or programming problem with creativity?

3 of 26

Overview

You should describe a research or data analysis question and explain its importance

You should summarize work that already exists (if it does)

You should outline the work you plan to do

You should demonstrate you have access to the data, describe the data, and propose how you will collect the data

You should describe the programming paradigms you plan to use and why it makes sense to combine them for your project

You should describe any packages and/or software you plan to use

You should briefly describe the data analytic product you plan to build

4 of 26

Background

Investing in stocks involves buying shares of a company to own a portion of it

You earn a profit when stock price goes up�

Retail investors: individual, non-professional investors that buy and sell for their personal accounts�
Institutional investors are large organizations (e.g. hedge funds, banks, insurance companies)

5 of 26

Project Objective

S&P500 2023 performance (%): 26.3�Hedge funds:

Data from https://fundselectorasia.com/10-funds-that-beat-the-sp-500-by-over-20-in-2023/

Unfortunately, it is very hard for most retail investors to make successful individual stock picks. To put this into context, in 2023, the S&P500 had an annual return of 26%. Top hedge funds had upwards of 50%. However, on the right, we see that the average retail investor consistently underperforms the S&P500.

So why do we do so poorly compared to institutions? Firstly, most people don’t have all day to do research into market trends. More importantly, extensive and proprietary access to resources, data, expertise, and lower transaction costs give institutions a competitive edge.

For most of us, the easiest way to maximize returns and reduce risk is to invest in index funds. However, simple adjustments can be made to improve our returns. Here, we create a tool to help retail investors make informed decisions to optimize their portfolio.

Source: https://fundselectorasia.com/10-funds-that-beat-the-sp-500-by-over-20-in-2023/

6 of 26

Existing work

Gold-standard: Bloomberg Terminal

$30-50k/year�

Free options: Yahoo Finance/Stocks tracker

Real-time market data, news, portfolio tracking�

Jargon-heavy, investment research tools that can be confusing and distracting for many (and still very time consuming)�
Lack of statistical modelling and prediction

Stock Tracker.com

First, I’d like to preface that these tools exist. The Bloomberg Terminal, for example, is a powerful financial software platform that provides real-time data, analytics, news, and research documents. The caveat is that a subscription costs $30,000-$50,000 per year.

Free platforms also exist.

But, all of these may be jargon-heavy and provide way too much information to sift through – most of which will not be relevant. The free tools also do not provide any quantitative modeling or prediction capabilities.

On the other hand, our website is targeted to recreational investors with limited time and industry knowledge.

REMOVE

Our features focus on helping the user optimize their portfolio with respect to the S&P500. We also provide some statistical tools, such as monte carlo simulations to forecast your portfolio’s performance.

7 of 26

Programming Paradigms

Functional programming:

Higher-order functions such as sapply(), map() from the purrr package, e.g.:�`port_ret <- sum(map_df(price_data[-1] + 1, prod) * weights) - 1`
Anonymous functions, e.g. defining slider values in the UI for portfolio specification: �weights <- sapply(weight_id, function(id) {input[[id]]})
Function composition: helper functions were used in other functions�

Machine learning

Unsupervised learning: K-means clustering

8 of 26

Our Data

Data source: Yahoo! Finance + library(tidyquant)

9 of 26

Our Data

Stock price
Dividends
Split ratio
Economic data
…

10 of 26

Data Analytic Product

Shiny app: A tool that aims to give beginners an edge to better understand stock performances and optimize their investment strategy in the stock market.

Advantage:

Ease of use
Interactive data visualization
Customization tailored to users' specific needs
Future prediction with MCMC

11 of 26

Shiny App Structure

Introduction of this app: General introduction to the app, customize the theme
Data summary: General information about the stock market in a tabular format
Stock trends: Stock Comparison
Market Distribution: Market performance based on the S&P 500 and its sectors
Understand your portfolio: Customized portfolio performance using historical data and predictions through Monte Carlo simulations
Stock clustering: S&P 500 stock clusters based on return and volatility

1 of 26

2 of 26

3 of 26

4 of 26

5 of 26

6 of 26

7 of 26

8 of 26

9 of 26

10 of 26

11 of 26

12 of 26

13 of 26

14 of 26

15 of 26

16 of 26

17 of 26

18 of 26

19 of 26

20 of 26

21 of 26

22 of 26

23 of 26

24 of 26

25 of 26

26 of 26