# Workshop Wednesday: Engineering Star Trek with Data Science ## Product Requirement Documentation ### How to write PRDs [![prd](https://hackmd.io/_uploads/BJC_qmMLxx.png)](https://www.aakashg.com/counter-intuitive-pm-moves/prd/) <small>image source: aakashg.com</small> - User Profile - Objectives and goals - Features (Essential / Stretch) - UX Flow and design notes - System and environment requirements - Assumptions constraints and dependencies - Timeline & Resources - Potential issues - Quality and standards ## Enterprise Bridge Computer (Jupyter + Streamlit) ## Starfleet Academy ML Predictor (scikit-learn) ## Universal Translator Neural Network (TensorFlow) ## 1. Tricorder Biological Diagnostics (SciKit-Image + ML) ### User Profile Average person who is not familiar with identifying plant species. Someone who has access to a mobile device (Android perfered, Apple ok) will take their phone with them on nature walks. ### Objectives & Goals Goal(User): Easy identification of plants with your mobile device. Goal(Programmer): Build a computervision model from scratch and learn how to create a application with BeeWare. ### Features (Essential / Stretch) - "Seen Plants" Object: Name, Date, Time, Picture, Description, Tags - Deletes "Seen Plant" - Add "Seen Plant" - Update "Seen Plant" - Add "Tag" - Delete "Tag" - Update "Tag" - Plant Diary: log of seen plants - "Plant Sighting" Object <---> one to many relation with "Seen Plant" - "Plants" Object: - Dashboard for benchmarking identification model ### Systems recs - NVIDA GPU - cuCIM - BeeWare - scikit-image ### UX Flow - User downloads application from github in an executable format - User is out in the world - Pulls out their phone (Android or Apple) and will take a picture - Use their phone's camera to take a picture of flowers - Use ComputerVision to identify plant species - Save it in personal catalog of plant species in application ## Holodeck Environment Generator (BlenderAPI, PyScript + NumPy) ## 2. Warp Core Optimizer (Numba + SciPy + DeepLearning) - Amazon SageMaker - PyTorch - NVIDIA GPUs - Numba v2 ## Stellar Cartography Mapper (HoloViz + 3D) ## Borg Pattern Recognition (PyTorch) ## Replicator Molecular Database (image processing) ## 3. Temporal Anomaly Detection (time series + DuckDB)