Watson Visual Recognition Workshop

Simple workshop using the classify, train & similarity search features of Watson Visual Recognition

Slides at https://ibm.box.com/v/assembly-watsonvr

This app can be used to run Python code interactively through Bluemix, for example, the app code demo-ed in this article: Ask Watson what Twitter is telling you: Part 3 - Analyze tweet pictures for categorization and recognition

Prerequisites

1. Python 2.7 [https://www.python.org/download/releases/2.7/], Pip [https://www.python.org/download/releases/2.7/]
2. Twitter app credentials [https://apps.twitter.com/]
3. Bluemix account & Watson Visual Recognition service credentials [https://www.ibm.com/watson/developercloud/doc/common/getting-started-credentials.html]

Files and folders

• Procfile: Contains the command that tells Bluemix which file to launch when the app is opened. For this app, the iPython framework is specified as the launchpoint.
• README.md: The file that you are reading at this moment.
• manifest.yml: A file that tells Bluemix a bit more about the app.
• requirements.txt: A file for listing the Python libraries that you need for your app code, with each library on a new line. Replace the libraries in this file with whatever you need for your app. For this app to run, the only required libraries are tweepy & watson_developer_cloud.
• getTweet.py: Retrieve tweets being tracked, analyze for emotional tone & store JSON output in tweetdata.json
• getImage.py: Extract Image content from tweet & store urls in images.txt
• getClassification.py: Classify Image urls using Visual Recognition & store JSON output in classifiedImages.json