JupyterCon2017

Notebook Narratives for Industry from 2017 JupyterCon

Sample notebooks for industry specific machine learning implementations

Designed for you to re-use for your own solutions

To run Jupyter Notebooks directly from this Github use Binder Turn a GitHub repo into a collection of interactive notebooks

Microsoft Cognitive API http://aka.ms/cognitive and Azure Jupyter Python Notebooks showcasing the Microsoft Cognitive APIs which can be used across curriculum.
As the pace of global innovation continues to accelerate, the University of Cambridge is evolving engineering & Mathematics curriculum to teach core concepts faster using higher level, open source tools in the public cloud. For example, a Dr Garth Wells has increased learning in an introductory computing class by having students use Microsoft Azure Notebooks, which allows them to spend more time mastering concepts and enhancing problem solving skills and less time on language syntax. This technology switch also gives students anytime, anywhere access to required tools needed to complete assignments, and it facilitates greater collaboration between professors, students, and the larger community. In addition, after Cambridge adopted a public cloud solution, IT infrastructure doesn’t limit the ingenuity of bright minds. Dr Garth Wells University of Cambridge - Mathematics & Enginnering Notebooks hosted on http://Notebooks.azure.com

Description - Querying specific content areas quickly and easily is a common services sector need. Fast traversal of specialized publications, customer support knowledge bases or document repositories allows service companies to deliver their particular service efficiently and effectively. Simple FAQs don’t cover enough ground, and a string search isn’t effective or efficient for those not familiar with the domain or the document set. Instead, these companies can deliver a custom search experience that saves their clients time and provides them better service through a question and answer format. In this project, we leveraged Azure Search and Cognitive Services and we share our custom code for iterative testing, measurement and indexer redeployment. In our solution, the customized search engine will form the foundation for delivering a question and answer experience in a specific domain area.
The notebook here allows you to pre-process and enrich source text, upload a search index, and interactively query that search engine.
Read the complete code story behind this example at the Microsoft Develop Blog here: https://www.microsoft.com/developerblog/2017/08/07/developing-a-custom-search-engine-for-an-expert-system/

Description - Using a photo of an end-cap, classify if the stocking is in-compliance or out of compliance with the planogram. Leverages CNTK. Uses Fast R-CNN.

Description - With wearable IoT sensors, we can collect positional and motion data that allow us to measure this expertise level distinction between professionals and amateurs with high precision and accuracy. In our analysis, we discovered the sensor data from just nine body positions provides ample signal to generate distinct activity signatures for the professional skiers when compared with the amateurs. In the notebook we describe how we engineer features for time segments that describe differences that describe motion and expertise relative to the expert. This allows amateurs to analyze deficit areas to improve their skiing.

catalystcode / jupytercon2017 Goto Github PK