Weather is a source of fascination and endless conversation. When you combine weather data sets with twitter chatter about the weather, what kinds of surprising insights might you find? Are the weather sentiments on Twitter related to the actual weather and forecasts? Or is there something else at play? This talk shows you how to extract, combine, and analyze Twitter and weather feeds in Python, using Spark, Insight for Twitter, Weather Company Data for Bluemix and Watson Tone Analyser. We will explore how to analyse this data in a Python notebook to draw insights into patterns and sentiment that may not otherwise be apparent.
The talk covers how to:
– Extract Twitter feeds daily and store them in a Cloudant database.
– Analyze the tweet sentiment using Watson Tone Analyzer. – Deploy Weather Company Data for Bluemix to collect daily weather data and load it into Cloudant.
– Combine and analyze the stored data to understand correlations between sentiment and weather in a Python notebook using Spark and multivariate statistical methods. We’ll also use clustering algorithms to compare correlations between regions to find out if the English weather really is as bad as they say.