At a hackathon you have a limited amount of time to come up with an idea and build a solution. This post will show you different features in Microsoft Azure along with links to tutorials to help you build an awesome hack no matter what programming language or Operating System you choose!
This post will cover
- Analyzing images
- Analyzing video
- Analyzing speech
- Analyzing text
- Training your own data models
- Internet of Things
- Storing data
- Compute power
- Web site and web service hosting
- After the hackathon
If Microsoft is at your hackathon, stop by their booth to see if they have any Azure passes available. If not, you can get a free trial.
There are cameras everywhere: web cams; phone cameras; drones; dash cams, security cameras. What could you do if you analyzed the images from those devices? Cognitive Services has a series of Visionfires APIs and tools to help. You can detect emotions; recognize faces; Build a custom image classifier; Content moderation to prevent inappropriate content; read text in images; generate thumbnails. At hackMIT in Boston students trained a custom image classifier so they could use a Roomba to find lost objects.
You are not limited to analyzing still images, you can analyze video as well. Cognitive Services includes Vision APIs for videos: Face detection and tracking; Motion detection; Stabilization for shaky videos; Generate video thumbnails.
Use the Video Indexer to get a transcript of the spoken words (9 languages supported and more to come); map and understand which speaker spoke and when; extract text displayed in the videos; separate background noise and voice activity; determine when a scene changes in the video; sentiment analysis; translate the audio transcript from one language to another; moderate inappropriate content; extract keywords.
Recently at a hackathon some students built a solution to analyze the reactions of people watching videos to help advertisers learn the effect of their commercials.
More and more devices are accepting voice commands. Controlling anything by voice requires Speech APIs. I highly recommend using an external microphone when you test and demonstrate any project using speech to eliminate as much background noise as possible. Microsoft provides Speaker recognition; Speech to text; Text to speech; Translation to and from 10 languages; Custom Speech Service to create customized language models to handle specific jargon that may not be recognized by everyday speech to text models. Could you build an app to help people visiting countries where they do not speak the native language?
The world is full of tweets, articles, emails, text books, research papers, and more! How can we create and analyze their content more efficiently? Spell check; Sentiment analysis; Key Phrase extraction; Language Detection; Translate between languages; Crowd source translation improvement. Could you use any of these features to detect fake news?
Training your own data models
Do you have data you want to analyze to find trends? You can prepare your data, build and deploy models with Azure Machine Learning ; Or use a deep-learning tool such as Microsoft Cognitive Toolkit installed on a virtual machine running on Azure. Check out the Data Science Virtual machine to save time installing software. A student at a hackathon in Montreal, Canada interviewed other students to collect data and trained a model to determine how many hours you need to study to pass an exam.
Web site and web service hosting
Whether you developed your application with Python, Java, PHP, Node.js, or .NET you can deploy that app to Azure.
After the hackathon
Did you win at the hackathon? Maybe you came up with a cool idea but did not have time to finish before judging? Check out the Microsoft Imagine Cup our worldwide student technology competition. Keep working on your hack and it could take you all the way to the world finals.