# Context I recently implemented an iPhone App for my younger daughter to learn multiplication. It probably was as much of a learning exercise for me as it was for her. But now I want to tackle the next challenge and add some more smarts with a module to recognise her handwritten digits. Instead of taking a pixel based approach, I took the opportunity of the touch-based input device to record the training data as vectorized paths. # Content The data contains 400 paths for digits along with their matching labels. Each path is normalized to 20 vectors (2d) and the total length of the vectors is normalized to ~100. Find an example record of the JSON-structure below: { "Items": [ {"y":1, "id": 1488036701781, "p": [[0,0],[1,-2],[7,-9],[3,-5],[6,-9],[2,-3],[3,-4],[1,-1],[1,-1],[0,0],[0,2],[0,10],[0,8],[0,17],[0,6],[0,10],[0,3],[0,3],[0,0],[-1,1]] }] } # Acknowledgements Thanks to my daughter for the training data! # Inspiration I hope you have fun with learning and I very much welcome hints on how to better capture the data.