+41 44 500 21 28

Do. Data with Quantum

Sentiment Analysis KNIME

Build your own sentiment analysis step by step

Do you want to know what your customers/users/contacts/relatives really think?

You could actually build your own sentiment analysis application. In this workshop, we will work together to build such an application step by step. That is: data access, text cleaning, stemming, dictionary tagging, text visualization and machine learning techniques.

We are also planning to reserve the last 15 minutes of the workshop for a Q&A session with Vincenzo and Rosaria, the authors of the book on text mining From Words To Wisdom. So, if you have a burning question about text mining or if you get one while working on the sentiment analysis application during the workshop, here is your chance to ask the experts!

When: Thursday, 21 Jun 2018, 16:45 – 19:15+
Where: Technoparkstrasse 1, 8005 Zürich / Conference room name: Fortran
Cost: Free

>> Sign up now! We look forward to seeing you there!

This event is co-organized with the Switzerland KNIME User meetup group.

Agenda
16:45 – 17:00 Arrival and registration
17:00 – 17:10 Welcome by Quantum
17:10 – 17:30 Intro to KNIME with demo
17:30 – 19:00 Hands-on workshop
19:00 – 19:15 Q&A with authors of the book From Words to Wisdom
19:15 + Apéro and networking

In order to get the most out of the hands-on workshop, you should bring your own laptop with the KNIME Analytics Platform pre-installed. However, it is also possible to work in groups sharing a laptop (a few laptops will also be made available).

To install KNIME Analytics Platform, follow the instructions provided in these YouTube videos: Windows | Mac | Linux

If you would like to become familiar with the KNIME Analytics Platform, you can explore the content of KNIME’s E-Learning course.

Our previous Do. Data events


AI & Image Recognition | 19 Apr 2018

During the event, we compared the photos of celebrities to those of terrorists to show how it’s possible for organizations such as the FBI to mistake someone’s identity if using only facial recognition detection. We also focused on how artificial intelligence can be applied by any organisation in the context of image recognition/classification and social media listening in order to constantly monitor users’ posts, comments and feedback (text or pictures), evaluate the sentiment, and react in a timely manner.

Contact us to understand how to quickly get started with artificial intelligence and image recognition within just a couple of weeks. We’ll show you how to leverage existing python packages, publicly available databases of images and other open-source resources. Or learn more about our Social Media Listening training course.

For more insights, check out our presentation. This presentation must not be copied, distributed or reproduced in whole or in part without Quantum’s written consent.


Artifical Intelligence AI Image Recognition


Big Data & Twitter | 19 Oct 2017

We compiled Twitter and relationship data on tweets from @realDonaldTrump from Twitter, Trump’s connections from BuzzFeed, and world’s richest people / most powerful people / top 2000 companies from Forbes lists. Whether it was storing or analyzing the data, the right tool was selected for the right job. Python to download the data from Twitter APIs. A document database (MongoDB) to store the JSON files. Tableau to analyze and visualize the Tweet patterns. Finally, a graph DB (neo4j) to analyze and visualize the connections/relationships. One insight was that Trump started to use paid marketing on Twitter at a time when his approval ratings dropped significantly, from visualizations of his tweeting patterns using Tableau. Another insight was based on a K-means clustering of Trump’s Twitter friends and followers. Surprisingly, not all of his followers were conservative-leaning, gun-rights protectionists. Rather, there were five distinct groups, including urbanites/professionals.

For more insights, check out our Tableau workbook below, download our presentation, and try out some of the Cypher queries in neo4j. Contact us if you’d like a copy of the Trump World neo4j database that includes all the data from Twitter, BuzzFeed, and Forbes lists.


Big Data Twitter


New York City Taxis & Uber | 18 May 2017

In New York City, with a population of roughly 8.5M, there are over 750K taxi rides every day (incl. Uber, Lyft, etc.). The market leaders are Yellow Taxis with 42% of the market, followed by Uber with 30% (as of 2017). With data downloaded from NYC OpenData/Taxi & Limo Commission, 25M records for Dec 2016 were visualized in Tableau. One surprising insight is Uber has a higher market share of rides in upper Manhattan and the outer boroughs and lower market share in lower Manhattan, likely because they face stiff competition from the Yellow Taxis in lower Manhattan (designated the “Yellow Zone” by the NYC Taxi & Limo Commission).

See our Tableau workbook below for the full story. Disclaimer: The data has been limited to Dec 16 – 31 due to Tableau Public limitation of 15M records.


Open Data New York Taxis Uber


Australia’s Gender-Pay Gap | 23 March 2017

The Australian Taxation Office published tax data for the years 2013-2014, with average income broken out by gender and occupation. Not surprisingly, males had higher average incomes than females for the same occupation, across almost all occupations. Exceptions included futures traders and mountain guides. While the data has some interesting insights, it doesn’t alone tell us if there’s a difference in the average hourly wage of females and males.

Peruse our Tableau workbook with different charts to inspire you. Charts include a boxplot, scatterplot, dumbbell (using female/male icons), and distribution.


gap analysis