READING HANDWRITTEN PIN CODES
=============================

A sample capstone project for CBSE Class XII, Artificial Intelligence
(Subject Code 843), Part C. From LambdaLab - https://www.lambdalab.in

WHAT IS IN HERE
  pin_reader.py               the program, one section per step of the Data Science Methodology
  pin_reader.ipynb            the same program as a Jupyter notebook
  digits.csv                  the data it reads
  pin_codes/                  the data it reads
  pin_truth.csv               the data it reads
  chart1_samples.png          a chart the program draws
  chart2_average_digits.png   a chart the program draws
  chart3_loss_curve.png       a chart the program draws
  chart4_confusion.png        a chart the program draws
  chart5_mistakes.png         a chart the program draws
  chart6_kfold.png            a chart the program draws
  chart7_pin_codes.png        a chart the program draws

HOW TO RUN IT
  1. Install the libraries once:  pip install pandas numpy matplotlib scikit-learn opencv-python
  2. Keep this folder together - the program looks for its data beside it.
  3. Either run:  python pin_reader.py
     or open pin_reader.ipynb in Jupyter Notebook and run the cells in order.
  The charts are saved into this folder as PNG files.

THE DATA
  The dataset is a real, public one. Its source and licence are given in the
  References section of the report. Say so in your own report too.

THE NO-CODE VERSION
  The curriculum also builds AI projects without code, in Orange Data Mining.
  The report shows how to build this same project there.

The full report, with every section CBSE marks, is on the project's page:
  https://www.lambdalab.in/learn/ai-12-projects/pin-code-reader
