AP Week and Big Project 3 Plans

[JishnuS420]

RDS and SQLite Legacy Challenge:

• For RDS, we want to have Adi, who has had prior knowledge on the topic working on it, and alongside Adi, we also have Ethan Zhao and Alex Lu helping him to complete the challenge

Big Project:

For the big project, a large majority of the backend is completed and the key features are primarily done apart from things for the leaderboard and other aparts of the stock game and so, the frontend is the primary focus now.

Links to individual issues are highlighted in blue.

• David <3 will be moving over to Frontend but primarily to just correct things and fine tune them to the backend

• Jishnu will be countinuing to improve both graphs for the live stats and also the LSTM prediction and will be working to add someother features to the live stats and will work with David to fix it up

• Emaad will be fixing up the styling page wide and also making sure that the pages are consisstent amongst all the pages and have a proper working global navbar and will work with all the frontend components

• Ethan Tran and Krishiv will be working to fix up the leaderboard things and work on making it work on both backend and frontend, working alongside James

• Anthony will be working on the backend fixing up the perceptron model and making sure that model is making it more accurate and will be working with Tay

• Tay is backing up Anthony, also working on the backend making the model more accurate and also working on the frontend of the perceptron model

• James is working on the frontend for the stock game and will be working along with Ethan and Krishiv to make the stock game close to completion

• Yuri will be finishing up the homepage and making it fluid and smooth as well as work on a little bit of blogging

Features

Ethan T

Our backend has...

• [x] A working AI model that makes predictions on what sign the user is making with their hands (ASL Alphabet and Custom signs)
• [x] Data storage for each frame from the backend (what we use to make predictions)

Hand Recognition Generalization

Improve accuracy of the model (further training of the model for accuracy) OpenCV's bounding box capabilities rely solely on geometric properties and the pixels taken from the frontend in the bounded area. Through Generalization, the hand's appearance in different environments such as lighting.

Timeline

• [x] Research generalization, understand how it works and why it's implementation will allow for adaptability in the machine learning model
• [x] Code for Data Augmentation of original frames I am working on code that takes each frame and creates new data based off the original frame in order to account for variability. This would allow the model to be trained on each frame in different lighting scenarios and allow for an accurate generalization/prediction.

    // method to create variations of an image
    private BufferedImage[] createAugmentedImages(BufferedImage originalImage) throws IOException {
        BufferedImage[] augmentedImages = new BufferedImage[9]; // array size remains 9 for 9 variations
        augmentedImages[0] = originalImage; // first image remains unchanged
        augmentedImages[1] = changeBrightness(originalImage, 0.8f); // decrease brightness by 20%
        augmentedImages[2] = changeBrightness(originalImage, 1.2f); // increase brightness by 20%
        augmentedImages[3] = addNoise(originalImage, 0.1); // add low level of noise
        augmentedImages[4] = addNoise(originalImage, 0.2); // add higher level of noise
        augmentedImages[5] = adjustHue(originalImage); // randomly adjust hue
        augmentedImages[6] = rotateImage(originalImage, 30); // rotate image by 30 degrees
        augmentedImages[7] = flipImage(originalImage, true); // flip image horizontally
        augmentedImages[8] = flipImage(originalImage, false); // flip image vertically

        // save each augmented image to a file using i+1 as filename
        for (int i = 0; i < augmentedImages.length; i++) {
            saveImage(augmentedImages[i], "augmented_image_" + (i + 1) + ".png");
        }

        return augmentedImages;
    }    

    // method to change brightness of an image
    private BufferedImage changeBrightness(BufferedImage image, float factor) {
        BufferedImage newImage = new BufferedImage(image.getWidth(), image.getHeight(), image.getType());
        Graphics2D g = newImage.createGraphics();
        g.drawImage(image, 0, 0, null);
        g.dispose();

        // adjust brightness for each pixel
        for (int y = 0; y < newImage.getHeight(); y++) {
            for (int x = 0; x < newImage.getWidth(); x++) {
                int rgba = newImage.getRGB(x, y);
                int alpha = (rgba >> 24) & 0xff;
                int red = (int)Math.min(255, ((rgba >> 16) & 0xff) * factor);
                int green = (int)Math.min(255, ((rgba >> 8) & 0xff) * factor);
                int blue = (int)Math.min(255, (rgba & 0xff) * factor);
                newImage.setRGB(x, y, (alpha << 24) | (red << 16) | (green << 8) | blue);
            }
        }
        return newImage;
    }

    // method to add noise to an image
    private BufferedImage addNoise(BufferedImage image, double noiseLevel) {
        BufferedImage newImage = new BufferedImage(image.getWidth(), image.getHeight(), image.getType());
        Graphics2D g = newImage.createGraphics();
        g.drawImage(image, 0, 0, null);
        g.dispose();

        // add noise to each pixel
        for (int y = 0; y < newImage.getHeight(); y++) {
            for (int x = 0; x < newImage.getWidth(); x++) {
                int rgba = newImage.getRGB(x, y);
                int alpha = (rgba >> 24) & 0xff;
                int red = (int)Math.min(255, Math.max(0, (int)(((rgba >> 16) & 0xff) * (1 + (Math.random() - 0.5) * noiseLevel))));
                int green = (int)Math.min(255, Math.max(0, (int)(((rgba >> 8) & 0xff) * (1 + (Math.random() - 0.5) * noiseLevel))));
                int blue = (int)Math.min(255, Math.max(0, (int)((rgba & 0xff) * (1 + (Math.random() - 0.5) * noiseLevel))));
                newImage.setRGB(x, y, (alpha << 24) | (red << 16) | (green << 8) | blue);
            }
        }
        return newImage;
    }

    // method to adjust hue of an image
    private BufferedImage adjustHue(BufferedImage image) {
        BufferedImage newImage = new BufferedImage(image.getWidth(), image.getHeight(), image.getType());
        Graphics2D g = newImage.createGraphics();
        g.drawImage(image, 0, 0, null);
        g.dispose();

        // adjust hue for each pixel
        for (int y = 0; y < newImage.getHeight(); y++) {
            for (int x = 0; x < newImage.getWidth(); x++) {
                int rgba = newImage.getRGB(x, y);
                int alpha = (rgba >> 24) & 0xff;
                int red = Math.min(255, (int)(((rgba >> 16) & 0xff) * (0.8 + Math.random() * 0.4)));
                int green = Math.min(255, (int)(((rgba >> 8) & 0xff) * (0.8 + Math.random() * 0.4)));
                int blue = Math.min(255, (int)((rgba & 0xff) * (0.8 + Math.random() * 0.4)));
                newImage.setRGB(x, y, (alpha << 24) | (red << 16) | (green << 8) | blue);
            }
        }
        return newImage;
    }

    // method to rotate an image
    private BufferedImage rotateImage(BufferedImage image, double angle) {
        // create a transformation matrix for rotating the image
        AffineTransform tx = new AffineTransform();
        tx.rotate(Math.toRadians(angle), image.getWidth() / 2.0, image.getHeight() / 2.0);

        // apply the transformation to the image
        AffineTransformOp op = new AffineTransformOp(tx, AffineTransformOp.TYPE_BILINEAR);
        return op.filter(image, null);
    }

    // method to flip an image horizontally or vertically
    private BufferedImage flipImage(BufferedImage image, boolean horizontal) {
        // create a transformation matrix for flipping the image
        AffineTransform tx = new AffineTransform();
        if (horizontal) {
            tx.scale(-1, 1); // flip horizontally
            tx.translate(-image.getWidth(), 0);
        } else {
            tx.scale(1, -1); // flip vertically
            tx.translate(0, -image.getHeight());
        }

        // apply the transformation to the image
        AffineTransformOp op = new AffineTransformOp(tx, AffineTransformOp.TYPE_NEAREST_NEIGHBOR);
        return op.filter(image, null);
    }

• [x] Implement Generalization to the model itself After completing the code that will take into account the variations ie. lighting, horizontal, tilt and form the next step is to utilize the mnist data from the model and use this in the model which will make the predictions

    private List<List<List<Integer>>> generateAugmentedFrames(String imageData) throws IOException {
        byte[] imageBytes = Base64.getDecoder().decode(imageData);
        ByteArrayInputStream bis = new ByteArrayInputStream(imageBytes);
        BufferedImage originalImage = ImageIO.read(bis);

        List<List<List<Integer>>> mnistDataList = new ArrayList<>();
        BufferedImage[] augmentedImages = createAugmentedImages(originalImage);

        for (BufferedImage image : augmentedImages) {
            mnistDataList.add(convertToMNIST(image));
        }

        return mnistDataList;
    }

Krishiv

LeaderBoard Key Features

---

Backend

• Object: Declare variables and instantiate them and create columns for the SQL DB in order to have raw data to be able to sort specific users based on account growth and give them a rating.
• Collectable: Used as an abstract class( a class which isn't instantiated on its own) to set key types and variables for the performance, here we also initialize the specific variables.
• Itterator: The iterator manages the object through an interface for the performance object and as an iterator the function of the file is to The class includes methods to get the list size, update the key type for sorting purposes, and perform a merge sort on the list. This is to sort by the key value of account value that was set in the collectable and make the full dataset be active and work.

Working Feature

USE OF SORTING IN THE ITTERATOR

SQL IMPLEMENTATION

POSTMAN REQUESTS

FRONTEND FETCH AND WORKING TABLE!

ERRORS THAT WERE FACED!

1. POM.XML; with working on different versions, since everyones codebase was constantly changing using wrong headers and imports often left dependencies to fail and build not working in the front-end, however this was fixed through better communication and better branch management.

2. Being able to have test data: every time we tried to make a postman request to post new data and later on make a get request so that it sorted the list by a specified attribute, we would get some sort of error, such as a 405 Not Allowed error or even an Internal Server Error. Later on, however, we found that this was because we were missing PostMapping in our API controller for the performance leaderboard, so we added that along with a new endpoint so that it could create new users. Now we are able to add users to the list along with their other attributes.

(used the wrong endpoint and format when adding the data)

Jishnu

Things to tend to:

Live Stats page (primary priority):

• [x] Stats live that cleans up the graph and then updates to the tike range selected (1D, 1W, 1M, 1Y, Alltime)
• [x] Stats live has a carousel slider that displays similar stocks and popular stocks, Example:
• [x] Stats for a company is pulled and then also provides detail displaying the recommendations, Example:
• [x] Stats also shows the popular, gainers, and other relevant information

Prediction page (primary priority):

• [ ] Fix up the date scaling (change the graph to an irregular timeseries) example below:

Stock Game page (secondary priority, cannot be implmented without the above):

• [x] Implement the live graph into the game
• [x] Have the predicitons show up with the player's performance
• [x] The player's predcition are compared with the AI's predictions, and the real prices, award points accordingly

Timeline and CheckPoints:

• [x] Tuesday 5/21, check in with David after completed stats live and start steady implementation into the game
• [x] Thursday 5/23, check in with David for prediciton implementation and further work to be done

[akshat122805]

Team Features

• [ ] Login and sign up feature, however was not working initially
• [x] User authentication via photo ID, using filters to confirm the initial user
• [x] Trading game dashboard -- looks very nice. HOWEVER some features weren't working on the dashboard.
• [x] Get stock information works well, overall the project looks nice, especially this feature
• [x] Using 2-3 APIs, AlphaVantage for graphs and FinHub for stock information
• [x] Uses PyTorch library to analyze stock graphs, utilizes AI in multiple formats
• [x] Utilizes both python and java for implementing artificial intelligence
• [x] Utilizes sorting methods to organize stock tickers based on filters like growth equity, market cap, earnings, volume, etc.
• [ ] Leaderboard feature
• [x] Utilized stock iterator that uses merge sort as well as arraylists
• [ ] Fixing stats page

Individal Features

• [x] Ethan T: Login Authenticator
• [x] David: General Dashboard
• [x] Jishnu: Stock Information Dashboard (graphs, news, recommendations)
• [x] Alex: AI implementation with both python and java.
• [x] Emaad: Stock ticker dashboard that shows technical financial atrributes/indicators
• [x] Ethan Z.: Worked on RDS Feature/Blog
• [x] Krishiv: Backend for iterator
• [ ] Yuri: n/a

Takeaways

• Slow start, a lot of the features were not working initially, but they were able to gain their traction as the presentation went on.
• There was a big interruption, a technical issue lasted 3-4 minutes, which should have been practiced
• Alex did a great job validasting his feature, answered 3-4 questions diligently and showed complete understanding of his feature.
• Liked how Emaad utilized sorting for organizing stock tickers, reminds me of using stock screeners like FinViz
• The overall project looks great, good work on the CSS styling
• Some features were great in my opinion, like Alex's use of AI to predict the stock for the next few years or Jishnu's main stock information hub.
• Other features I didn't really understand, for instance the photo capture
• Overall the work was great, but presentation could have been much better

[rohinsood]

Ethan Z

• Did a well-planned intro about the project
• RDS blogging, created a template project

  David

• Deployed on personal nuc to account for the computing and space requirement of the project
• Login & Sign up feature
  • Not working, frontend was clean

    Ethan T

• Image recognition feature for gang signs, prediction to the user, displayed use through JWT
• Needed more code explanations

  Jishnu

• Stock analytics, graph generation
  • API usage for graphs, stock news, & stock recommendation percentages

    Alex

• AI prediction, python & java
  • Java LSTM, deprecated
  • Switched into python
  • Verified functinality in PostMan
  • Had RMSE for performace metrics
• Created a custom collectible for the StockObject
  • Has a built-in merge sort feature

    Emaad

• Stock Ticker feature
• Use sorting by whatever the key is, api endpoint

  Krishiv

• Built a collectable for performance metrics

[VishnuAravind12]

David

• Stunnning display of the JWT login feature and sign up feature.
• But the problem was the sign in and up was not very functional.
• Wow, it was deployed on their own server, that s creative, especially cuz one needs better computing resources for these higher level models.
• Stcok dashboard seemed non functional, quite a shame.

Ethan

• Gang sign biometric login is very creative, not something I have seen to often.
• Seemed functional.
• Am more interested in the workings of the CNN though, Would have loved to see an architecture schematic or the source code.

Jishnu

• Research stock feature seemed really nice to look at, UI wise.
• Fully functional as well, I see a lot of value in this feature, graph seems really easy to interact with.
• Stats seem great as well, and the stock news stuff seemed really cool.
• Use of multiple APIs to get all this information is really impressive.

Alex

• Very impressive feature, seems like this could be useful
• One in python and one in java is very impressive
• I have also have had trouble with DeepLearning4J to do anything useful
• The graphs seem really accurate
• Maybe a dropout layer would have been useful in curbing some of that overfitting
• Collectable and iterator objects for the stock data is also a great use of Java features

Emaad

• Presented the stats for the different companies
• Growth, open, high, low, volume
• Seems pretty useful, nice implementation of OOP to create a stock object
• Seemed like the user stats were not too functional unfortunately

Krishiv

• Iterator and Collectable for user stock stats, like acount value
• Makes for a veyr creatively implemented leaderboard
• Did not demo this?

Ethan Z.

• I liked his documentation with regards to the AWS stuff
• The RDS and SQLite implemetation was very impressive

Takeaways

• Too many feature errors
• But the features that did work were stunning

[vardaansinha]

1. David Vasilev

• Showed splendid knowledge of the register and login system!! He knows JWT..
• Troubleshooted presentation issues really well.

2. Ethan Tran

• The gang signs on the screen, showed and explained "capture image" feature really well and went step by step to explain biometric login system (both backend and frontend)
• It would've been nice to show the actual code behind it but I didn't really miss it because the explanation was great

3. Jishnu

• Nice, engaging lead in to explain the historical stock data feature
• Demonstrates all parts of the feature really well
• Explains what API they use, how they pull from the API, formatting the API data in a readable, visual format

4. Alex L

• Explained the thought process and planning, showed key progress and learnings over time (DeepLearning4J)
• Transferred over to PyTorch, showed what's really going on behind the scenes with the historical data and how the LSTM works
• Answered all questions really well, had demos to backup his answers and showed a lot of key knowledge of the architecture of the training model
• Made a stock object iterator implementing merge sort algorithm with set key type in order to determine the specific key that they want the array to be sorted with

5. Krishiv

• Used sorting based on account value --> key is the account value, you can get the account values and then you get a letter from a-d, within the DB they compare the percentage of growth of the accounts, get a rating based on that
• Showed the backend code and architecture well, explained both sides

6. Ethan Z

• Worked on RDS & SQLite implementation (challenge requirement)
• Showed blog that explained setting up RDS, shows collaboration aspect

TEAM:

1. Deployment, a full deployment of frontend and backend

   • [x] Key features must be deployed in order to be presented, localhost presentations are .55
   • [x] Multiple people need to be able to interact with features at the same time, data updates

2. Searching/Sorting, 2 features must support searching and sorting from backend,

   • [x] 1 feature should have JPA build in sorting,
   • [x] 1 features should perform custom sort after retrieving backend data
   • [x] Extra. A feature that utilizes Stacks or Queues for ordering

[Tirth-Thakkar]

David:

• JWT Login system with name, email, password, and dob, and Login
• Backend local machine login system and Route 53 routing to host backend with more compute power and storage compared to AWS alternatives to circumvent issues with AWS setup system

Alex:

• Two variants of the program in both Java and Python
• With DeepLearning4J however, due to the LSTM function being poor switched over to Python with Pytorch and better prediction capacity and got a better accuracy for the model
• The model will overfit due to the past performance to extract one further date and through the usage of stack information to allow for a bank of predictions
• Inspired by the prior Builder design pattern and adding a sequential layer functionality and well-designed architecture
• Good logging of messages and information during the training process and performance data
• Made Custom iterator and collectible objects and iterator patterns that allow for the custom sorting methods to be able to sort the Stock objects

Jishnu:

• Stock trading and management portal to display collected and managed data
• Well-created graph and data representation to show the prices in stock price changing
• Showing interesting stock recommendation information along with articles that are covered about the topics
• Good mix of a variety of APIs to pull information about the topics presented

Anthony:

• Biometric sign-in system with log-in system with image recognition, and variants of the image with increased and decreased lighting, noise adjustment, etc. To allow for more possibilities in the image that can occur in day-to-day situations.
• With the ability to detect and recognize faces and login into the portal

Emaad:

• Global styling changes and general UI revisions
• Created a Stock dashboard with Stock objects in a table demonstrating the various attributes of the Stock object with the main demonstration of sorting by the desired key value and sorting by the UI desired attribute
• Allowing for a user to consider what to trade in and use to be able to make a reasonable decision on purchasing
• Attributes are defined as fields and the creation of the needed endpoints
• Good usage of custom sorting algorithms in the sorting process from the given key that is passed by the table
• Showcasing of the account rating system and the general stock performance and account value in a leaderboard function

Ethan

• RDS Amazon Database setup and usage implementation along with a thorough blog explaining the setup and implementation of the functionality for Amazon
• Good explanation of Strengths and weaknesses of the features that are presented and the differing variants that can be used
• Username and advice provided for RDS setup and good usage for security and planning on RDS and setup plans
• Usage of RDS to store data updates functionality

Krishiv:

• Setting the key to an iterable object and sorting using the performance model and db calls
• Use the iterator pattern w/ merge sort and account value to create a leaderboard functionality and a letter that is returned A-D and shows the percent growth for the account and the specific rating that is provided

James

• N/A

Yuri

• N/A

Tay

• N/A

Presentation

• Multiple computers not present
• Issues with connection
• Issues with some features not working properly and errors
• Issues with certain demos not showing properly and smoothly based on the work that was created
• More general practice recommended

[h4seeb-cmd]

David

• JWT Login
• Custom deployment server on laptop using port forwarding and Route 53. Needed more computing power and storage than AWS could offer. Thus, this team decided to just host their own. Good solution to problem.

Jishnu:

• Live statistics page
• Page which displays information about a given stock so you know what you're investing in Feedback:
• Great use of APIs. The addition of the reccomendations page which lets you know what you're getting into is super creative.

Alex:

• Two different backends, one in java and one in python.
• DeepLearning4J used java-side.
• Pytorch used python-side, which phased out the java backend, as the LSTM function performed poorly java-side.
• The model takes the stock's performance in the last 60 days to figure out what the next data value is going to be.

Emaad:

• Responsible for all UX
• CSS work very pleasing to the eye
• Jank found within buttons and such can be improved on. Sites redirected to were 404'ing
• Created Stock Dashboard
• Many relevant data points found in dashboard
• Uses custom iterator to move through stock objects.
• Uses merge sort to sort the stocks in the dashboard by the desired stat, which can be toggled.
• Showcases leaderboard function, data not sorted/cleaned though...

Krishiv:

• Backend for the leaderboard
• Uses merge sort to sort through users to find account.

Ethan Tran:

• Face ID Biometric login system.
• Interesting addition, haven't seen another team do this.
• Works fairly well, is functional.

Ethan Zhao:

• Set up RDS
• Extremely in-depth documentation on RDS and how to set it up.
• Both python and java use RDS in order to store data.

Presentation Grows:

• Janky site functionality, especially with redirects
• Multiple computers not being utilized in Ditto
• Needed more practice in presentation

[Henerystone]

David login and sign up feature. problems with sign in functionality interesting application of creativity with the server but dashboard did not work

Ethan biometric login seemed interesting and everything seemed functional

Jishnu Research stock had nice UI. Fully functional graph seems easy to use. Use of multiple API

Alex interesting feature dual language usage is interesting too interesting graphing

Emaad Created a Stock dashboard, good use of CSS and staying looks good

Krishiv Iterator and Collectable for user ratings not able demo this?

Ethan Z. AWS documentation The RDS implemetation

Takeaways showcase of too many features, lower the amount and set aside ones that don't work

[Pitsco]

Giggle Juice's Reflection based on Kaiden, Derrick, and Nikhil's review

Notes

• Trade similar stocks → recreate Investopedia
• Implement some AI into it
• Role add student, add teacher and admin
• Different sessions: CSA-AI, Math101
• Purchase History, AAPL example stock data (real-time) pulled from an API (i’d assume)
• Rapid API
• It shows the live graph as well and can buy and sell shares
• Paper Money
• The stock ai thing was run on localhost but the hand sign thing was on the deploy page
• Also the sign thing to make a password for the person
• Feature: ai recognition thing of person or hand (not working, size too big)y
• Planning to use it for something like 2 factor authentication
• The AWS stuff doesn’t work on the school wifi
• Hand/image recognition works based on depth recognition.
• Can change the profile picture
• Perceptrion was tay and anthony
• RDS and some of the endpoints was adi
• Frontend was James

Final Grade: 11/12