Architecture Diagram
Image Sources:
https://upload.wikimedia.org/wikipedia/commons/1/1f/Python_logo_01.svg
https://upload.wikimedia.org/wikipedia/commons/8/83/AnantP%40javascript.png
https://upload.wikimedia.org/wikipedia/commons/thumb/8/8e/Nextjs-logo.svg/207px-Nextjs-logo.svg.png
https://user-images.githubusercontent.com/6509926/70553550-f033b980-1b40-11ea-9192-759b3b1053b3.png
https://a0.awsstatic.com/libra-css/images/logos/aws_logo_smile_1200x630.png
Platform Selection
| Technologies | Pros | Cons |
| JavaScript/TypeScript ✅ | -Used commonly for web development so necessary for our web app -Supports many frontend frameworks that will make it easier to build functional components | -Very prone to bugs/issues as Javascript itself is a loosely-typed language, but by using Typescript rather than Javascript, which supports static/strong typing which will combat this issue and prevent bugs |
| Python ✅ | -Necessary for our app in order to run the NLTK Python library as well as other Python libraries such as pandas, numpy, and matplotlib -There exists multiple popular backend frameworks in Python such as Flask, Django, etc | |
| Next.js ✅ | -Open-source framework built on top of React, thus has all the added benefits of React as well as an easy migration from React to NextJS if needed -Provides server-side rendering capability by default unlike React to load web apps faster initially -Well-maintained documentation | -Less community support than the general ReactJS framework |
| Server-Side Rendering and a Web App ✅ | -NextJs supports Server-Side Rendering, thus our web app will be rendered accordingly -Server-Side Rendering renders much faster initially than Client-Side Rendering for dynamic websites | -Slower to load than Client-Side Rendering with apps that have many features/interactions which isn’t true for our application as of yet |
| FastAPI ✅ | -Allows for a much quicker, direct way to build an API in Python -Can be maintained and easily tested since FastAPI has a built-in data validation feature -Supports features that allow testing for API requests -Supports asynchronous requests | -Much newer compared to Flask which was brought out in 1999 -Has a smaller community support |
| Flask | -Popular open-sourced Python backend framework with a lot of support and -Considered to be more “light-weight” to use compared to other Python frameworks including Django, etc | -Does not provide a built-in data validation feature -Lacks support for asynchronous requests |
| AWS ✅ | -Client Recommended | -Complex -Takes a while to get configured and running compared to Firebase |
| Firebase | -Client also recommended -Multiple team members have familiarity with Firebase -Easy to use and understand with extensive documentation -Provides all services required for the app including dbms, storage, and hosting -Pay as you go service | -Rigid format in terms of reading/writing to database |
| DynamoDB ✅ | -DynamoDB is a NoSQL database used by AWS, since we will be be using AWS for both hosting and database, this will be our database | |
| S3 ✅ | -S3, a feature of AWS, is specifically designed to store files, thus this will be how we store any files we may or may not need to store | |
| Pandas ✅ | -Popular Python library in the data science world used to develop on datasets to create dataframes | |
| NumPy ✅ | -Widely-used and heavily supported Python library that supports mathematical functions allowing us to modify our dataframe, numpy is a dependency of pandas | |
| Matplotlib ✅ | -Extensive and well-supported Python library that will allow us to model and plot/graph our dataframes | |
| NLTK ✅ | -Popular Python library that will allow us to tokenize comments/titles/sentences and then run a sentiment analysis to get a score between [0,1] inclusive to determine the sentiment between negative to positive |
Specific hardware or OS being built for?: No.
No Client Restrictions
Sources:
https://radixweb.com/blog/nextjs-vs-react#next
https://scythe-studio.com/en/blog/client-side-vs-server-side-rendering
https://www.turing.com/kb/fastapi-vs-flask-a-detailed-comparison
https://www.codecademy.com/article/introduction-to-numpy-and-pandas
https://www.monocubed.com/blog/top-python-frameworks/
Access to reddit API and cleaning/wrangling reddit api data -> https://towardsdatascience.com/how-to-use-the-reddit-api-in-python-5e05ddfd1e5c