PhotoStat is intended to be a web-based AI to solve introductory hypothesis tests (z-tests and t-tests of Normal distributions) by inserting an image of a word problem or typing out the question.

This project is a work in progress. I will first implement the project without using machine learning. I will extract necessary data from the word problem using other methods for the initial implementation of the project. Then, I will re-implement the project using machine learning.

• Website with a simple interface to take/insert an image of a word problem or type out the question
• Option to take a picture of the word problem
• Option to upload a picture of the word problem
• Option to type out the word problem
• Allow users to view a step-by-step process of the hypothesis test
• Option to copy the solution to the clipboard or save it to a file
• View more information about the AI regarding its reliability (%) and the project

• Add a way to process the image into text
• Refine image to text processing accuracy
• Train a model to extract core data (e.g. mean, standard deviation, etc.) from the problem
• Train a model to perform the hypothesis test
• Train a model to differentiate between z-tests and t-tests
• Allow users to view a step-by-step process of the hypothesis test
• Allow users to copy the solution to the clipboard
• Allow users to save the solution to a file
• Allow users to share the solution with others
• Create a website for the project
• Connect the AI to the website
• Allow users to view more information about the AI regarding its reliability (%) and the project
• Create a histogram of the website's daily/weekly/monthly usage and save it to a spreadsheet

• Pytesseract
  • You will need to install:
    • pip install pytesseract
    • Download Tesseract at UB Mannheim
• Word2Number
  • pip install word2number

• Image to Text
• Monkey Learn Information on Text Analysis - resource for myself :)